Differential Effects of Exercise on fMRI of the Midbrain As- cending Arousal Network Nuclei in Myalgic Encephalomyelitis / Chronic Fatigue Syndrome (ME/CFS) and Gulf War Illness (GWI) in a Model of Postexertional Malaise (PEM)

Background: Myalgic Encephalomyelitis / Chronic Fatigue Syndrome (ME/CFS), Gulf War Illness (GWI) and control subjects had fMRI during difficult cognitive tests performed before and after submaximal exercise provocation (Washington 2020). Exercise caused increased activation in ME/CFS but decreased activation for GWI in the dorsal midbrain, left Rolandic operculum and right middle insula. Midbrain and isthmus nuclei participate in threat assessment, attention, cognition, mood, pain, sleep, and autonomic dysfunction. Methods: Activated midbrain nuclei were inferred by re-analysis of data from 31 control, 36 ME/CFS and 78 GWI subjects using a seed region approach and the Harvard Ascending Arousal Network. Results: Before exercise, control and GWI had greater activation during cognition than ME/CFS in left pedunculotegmental nucleus. Postexercise ME/CFS had greater activation than GWI for midline periaqueductal gray, dorsal and median raphe, and right midbrain reticular formation, parabrachial complex and locus coeruleus. The change between days (delta) was positive for ME/CFS but negative for GWI indicating reciprocal patterns of activation. Controls had no changes. Conclusions: Exercise caused opposite effects with increased activation in ME/CFS but decreased activation in GWI indicating different pathophysiological responses to exertion and mechanisms of disease. Midbrain and isthmus nuclei contribute to postexertional malaise in ME/CFS and GWI.

other changes induced by the exertional challenge. In a previous study, we examined blood oxygenation level dependent (BOLD) activation during a difficult, high cognitive load continuous 2-back working memory task and compared preexercise and postexercise scans. Control, ME/CFS and GWI were equivalent prior to exercise (baseline), but after exercise ME/CFS had a significant increase in blood oxygenation level dependent (BOLD) activation while GWI had a significant decrease in the dorsal midbrain, right middle insula and left Rolandic operculum [14]. The midbrain region of interest extended from the left to right periaqueductal gray (PAG) and to the adjacent right midbrain reticular formation (MRF), inferior colliculus and lateral lemniscus, and caudally to the right lateral isthmus ( Figure S1). Because these nuclei have profound influences on threat assessment, pain, negative emotion, attention, wakefulness, and instinctual neurobehaviours, it was of interest to assess the activation of relevant anatomical midbrain nuclei.
In this report, we re-analyze the original BOLD data [14] using a seed region approach to gain a preliminary understanding of the nuclei that were activated within the midbrain region of interest. The seed regions were selected from the ascending arousal network that was defined from histological sections and diffusion studies of brainstem white matter tracts [14]. BOLD signals in each of the target nuclei were assessed on preexercise and postexercise days.
The aim was to identify which midbrain nuclei were affected by exercise in ME/CFS and GWI subjects and to judge effect sizes to guide future confirmatory studies. We propose that affected nuclei may participate in the pathology of postexertional malaise.

Ethics
All subjects gave written informed consent to this protocol that was approved by the Georgetown University

Demographics
GWI, ME/CFS and healthy control subjects were recruited to this four day long in-patient study in the Clinical Research Unit of the Georgetown -Howard Universities Center for Clinical and Translational Science. Subjects had history and physical examinations to ensure their inclusion by meeting Chronic Multisymptom Illness [3] and Kansas [4] criteria for GWI, Fukuda [1] and Canadian [2] criteria for ME/CFS, confirmation of sedentary lifestyle for control subjects (less than 40 min of aerobic activity per week) and exclusion because of serious medical or psychiatric conditions such as psychosis [4,[15][16][17]. History of posttraumatic stress disorder (PTSD) [18] or depression [19] were not exclusions unless the subject had been hospitalized in the past 5 years. Subjects completed Chronic Fatigue Syndrome Symptom Severity [20], SF-36 quality of life [21], Chalder Fatigue [22] and McGill Pain [23] questionnaires, and had systemic hyperalgesia tested by dolorimetry [24,25].

Orthostatic postural tachycardia phenotypes
Subjects rested supine for 5 minutes with continuous EKG measurements and arm cuff blood pressure every minute. Average recumbent heart rate (HR) was calculated. Subjects stood up and maintained their posture for 5 minutes. EKG and blood pressure were recorded each minute. The differences between standing HR at each minute and average recumbent HR were calculated (∆HR). The procedure was performed at least twice before the first exercise, then 1, 3, 8 and 24 hr postexercise. ∆HR was used to define Orthostatic status [26,27].
(i) Postural orthostatic tachycardia (POTS) was defined by ∆HR ≥ 30 beats per minute at least 4 time points before exercise and during each postexercise measurement period.
(ii) Stress Test Activated Reversible Tachycardia (START) was defined by normal ∆HR before exercise, but at least 2 episodes with ∆HR ≥ 30 beats per minute after exercise. The phenomenon was transient as postural tachycardia returned to normal within 36 to 48 hr.
(iii) The normal postural response was defined as Stress Test Originated Phantom Perception (STOPP) based on original findings in GWI [30,31] where ∆HR was in the normal range of 12 ± 5 beats per minute and never exceeded 30 beats per minute.

Verbal working memory task
Subjects practiced the 0-back and 2-back working memory task in a mock scanner until they felt proficient [14,30,31]. In the scanner, subjects viewed an instruction panel stating 'REST' for 0.8 s followed by 19.2 s of a blank screen. The instruction '0-BACK' was viewed for 0.8 s followed by 1.2 s of a blank screen, and then a string of nine pseudorandomized letters (A, B, C, D) seen for 0.8 s each followed by 1.2 s of a blank screen per letter. Each time they saw a letter, subjects pressed the corresponding button on an MRI compatible fibre-optic four button box that was Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 23 November 2021 used with both hands. After a second 'REST' period, they saw the instruction '2-BACK' and again viewed a string of nine letters. They had to view and remember the first two letters, then press the button for the first letter when they saw the third letter (i.e. '2 back', 4 s delay). The 2-back task continued for seven responses. This cycle was repeated five times.

MRI data acquisition, preprocessing and analysis
All structural and functional MRI data were acquired on a Siemens 3 T Tim Trio scanner located within the Center for Functional and Molecular Imaging at Georgetown University Medical Center equipped using a transmit-receive body coil and a commercial 12-element head coil array as described previously [14,30,31]. Parameters for structural 3D T1-weighted magnetization-prepared rapid acquisition with gradient echo (MPRAGE) images were: TE =2.52 ms, TR =1900ms, TI = 900ms, flip angle = 9, FOV = 250mm, 176 slices, slice resolution = 1.0 mm and voxel size 1x1x1mm. Images were processed in SPM12 [32]. fMRI data consisted of T2*-weighted gradient-echo planar images (EPIs) acquired during the n-back tasks. EPI data acquisition parameters were: TR/TE = 2500/30 ms, flip angle = 90, FoV =205 mm 2 , matrix size =64x64, number of slices = 47, voxel size = 3.2 mm 3 (isotropic). Raw EPI data were preprocessed through the default pipeline within the CONN toolbox [33]. Briefly, steps were: (i) slice timing correction, (ii) subject motion estimation and correction, (iii) outlier detection for 'scrubbing' based on Artifact Detection Tools, (iv) co-registration with structural data, (v) segmentation and spatial normalization into standard Montreal Neurological Institute (MNI) space [34] and (vi) spatial smoothing with a stationary Gaussian filter of 6 mm full-width at half maximum (FWHM). Voxel size was 2.0 mm 3 (isotropic) after spatial normalization and conversion to Montreal Neurological Institute (MNI) space.
All within-subject and group-level image analyses were performed using the SPM12 software package [35]. After accounting for magnetic saturation by removing the first 6 scans, a paradigm based on the timing of events in the 2-back task ( Fig. 1) was applied to preprocessed EPI data to sort individual subject scans into instruction, fixation, 0back and 2-back bins. In the original analysis, one sample t-tests contrasted the BOLD signals from the 2-back and 0back scans of each subject and included estimates of the translation (x, y, and z) and rotation (roll, pitch, and yaw) as covariates of non-interest. The resulting 2-back>0-back contrast maps from every subject were sorted into the control, GWI, and ME/CFS groups [14].
For this seed region re-analysis, we used the Harvard Ascending Arousal Network atlas [36,37] in the Lead DBS software package [38-40] to define regions of interest (ROI) for midbrain nuclei. The mean BOLD signal for each ROI was extracted from each subject's contrast map, re-centered to a population grand mean of 0, and the normalized data analyzed in the MarsBaR 0.44 toolbox [41,42]. MarsBaR output was the BOLD activation levels for the 2-back>0-back contrast condition in each midbrain nucleus for the control, ME/CFS and GWI groups on the pre-and post-exercise days.  Figure 1) [36,37]. When viewed from the anterior (ventral) and posterior (dorsal) sides, the nuclei in the ascending arousal network suggest three layers in the coronal plane of MRI space that approximate their embryological origins [43][44][45]. Mesomere 1 contributes the superior (SC) and inferior colliculus (IC), midbrain reticular formation Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 23 November 2021 (MRF) and periaqueductal grey (PAG). The second layer is derived from rhombomeres 0 and 1 and includes the superior dorsal raphe (DR), pontis oralis (PO), and more lateral penduculotegmental nuclei (PTN, formerly pedunculopontine nuclei or PPN) [43][44][45]. The ventral tegmental area (VTA) is anterior when viewed at this level but extends from the isthmus to diencephalon. The most caudal layer had midline DR and median raphe (MR), and bilateral locus coeruleus (LC) and parabrachial complex (PBC) that are derived from rhombomeres 1 and 2. Embryological origins do not align with the coronal MRI projection because of the marked ventral flexion of the midbrain during development and distortion of the original neural tube structures. Nuclei with significant differences by BOLD were manually highlighted by red outlines using Procreate software.

Supplementary Materials
Supplementary materials contain partial correlation coefficient analysis and multivariate general linear modeling results. Appendix A contains demographics, questionnaire and fMRI data.

Demographics and questionnaires
As expected, there were more females in ME/CFS because of the known female predominance [6] and more male veterans with GWI from assault divisions deployed to the Persian Gulf (Table 1). PTSD was more common in GWI. Pressure induced pain sensitivity tested by dolorimetry was not different because male and female subjects were combined [25]. ME/CFS and GWI had comparable symptom scores that were significantly worse than controls with the exception of pain, SF36 Role Emotional and Mental Health that indicated more impairment in GWI than ME/CFS.
Preexercise BOLD values were assessed by self-reported demographics variables in a multivariable general linear model. The significant covariates were Orthostatic status, low back pain, depression, heart disease, gender and marital status (Table S1). The next iteration used the significant covariates as fixed factors and removed the other variables. Orthostatic status was the only variable to be significant (Table S2). and delta, positively correlated between Day 2 and delta, and negatively correlated for Day 1 vs delta (Tables S3-S6).
There were no significant correlations between subjective questionnaire data and objective preexercise or postexercise BOLD outcomes. The magnitudes of the significant correlations (p < 0.05 corrected) were low (R < 0.4 and R > -0.4 for the inversely scored SF36 domains,).

ANOVA
BOLD data were compared between groups defined by disease status (control, ME/CFS, GWI) and orthostatic status (START, STOPP, POTS) on the preexercise and postexercise study days. In the original study, dorsal midbrain activation was not different between groups before exercise [14].
Visual inspection suggested a trend for differences in BOLD between groups. Data from all seed region datapoints and subjects were contrasted between groups. Prior to exercise, ME/CFS had numerically lower BOLD (0.108 ± 0.032, mean ± 95%CI for n=504 datapoints = all regions of interest in all ME/CFS subjects) compared to control (0.297 ± 0.037, mean ± 95%CI, n=434 datapoints) and GWI (0.235 ± 0.024, mean ± 95%CI, n=1092 datapoints). This suggested reduced blood flow during the 2-back>0-back condition for ME/CFS at baseline.
In the ascending arousal network, control was significantly more activated than ME/CFS in bilateral PTN, L_PBC, and VTA, while GWI was higher than ME/CFS in L_PTN and L_PO ( Figure 1, Table 2).
The only difference based on Orthostatic status on Day 1 was greater L_PBC activation in POTS than STOPP (Table S7).
Males (0.249 ± 0.023, mean ± 95%CI) showed a trend towards higher BOLD activation than females (0.167 ± 0.028, p = 0.000012 by unpaired 2-tailed t-test) when all nodes and subjects were assessed. The difference was significant for L_PO (p = 0.046) (Table S8).  On the postexercise day, the relationship between groups became inverted as control (0.254 ± 0.035, mean ± 95%CI) and ME/CFS (0.260 ± 0.034) had significantly greater BOLD than GWI by ANOVA when all subjects and nodes were evaluated (p<10 -9 by 2-tailed unpaired t-test with Bonferroni correction). After exertion, control was higher than GWI for bilateral MRF, VTA and R_PTN, while ME/CFS was greater than GWI for midline PAG, DR and MR and right MRF, LC and PBC (Table 3) (Figure 2).
There were no significant differences based on Orthostatic status following exercise (Table S9).
Gender was not a significant covariate for any region postexercise (Table S10). However, there was a general trend for females (0.123 ± 0.028, mean ± 95%CI) to have lower BOLD than males (0.197 ± 0.021, p = 0.000032 by 2-tailed unpaired t-test) when all nodes and subjects were compared.  Table 3). GWI was significantly lower than ME/CFS in the R_MRF, PAG, DR, MR, R_PBC and R_ LC. There were no significant differences between the control and ME/CFS groups following exercise. Data were annotated as in Figure   1. ΔBOLD was positive in ME/CFS and negative in GWI indicating the significant dynamic effects caused by exercise in these two diseases (Figure 3). ME/CFS had higher increments than GWI for all seed regions except MRF and L_LC (Table 4). There were no differences between groups defined by Orthostatic status or gender.  Table 4).
Age, gender, PTSD, BMI and dolorimetry pressure thresholds were not significant covariates prior to exercise.
Postexercise mGLM evaluated the same fixed factors and independent variables. Disease status was significant after exercise. ME/CFS and Control had significantly higher BOLD activation than GWI in VTA, L_MRF and R_PTN (Table S14). Overall, ME/CFS was greater than GWI for all regions except L_PO, L_LC and L_PBC. More nodes were significant by mGLM than ANOVA (Table 3). Gender was significant for R_LC and R_PBC as males had greater BOLD activation than females after adjustment for the other variables (Table S15). Other significant interactions between disease, orthostatic and PTSD status, age and dolorimetry thresholds (kg) were detailed in the Supplementary Online Material (Tables S16-S18), but must be interpreted with caution because of concerns about the number of variables, sample sizes and potential overfitting of the data.
Incremental changes in BOLD between days (ΔBOLD) reinforced the differences between diseases found by ANOVA (Table 4). Estimated marginal means for Disease status bracketed zero for controls, were positive for ME/CFS and negative for GWI (Table S19).
Orthostatic status had a significant impact with higher activity for STOPP than START in R_LCΔ and bilateral PBCΔ ( Table 5). The 95% confidence intervals for POTS and STOPP bracketed zero indicating no change after exercise.
However, ΔBOLD was negative for the START group indicating a dynamic exercise-induced effect on brainstem activation in this phenotype. Effect sizes were small indicating that it may be difficult to reproduce the finding.

Light and sound sensitivity
Provocations with light and sound showed that sensory sensitivities were significantly worse in ME/CFS and GWI than control subjects before and after exercise (Figure 4). Frequency analysis of light sensitivity prior to exercise found scores of 0 or 1 out of 20 (no discomfort) in 83.3% of SC, 30.8% of ME/CFS and 21.1% of GWI. Exercise worsened light sensitivity in paired analysis for CFS (p = 2.7x10 -6 ) and GWI (p = 0.022), and sound sensitivity in ME/CFS (p = 0.037) by 2-tailed paired t-tests. The incremental changes (∆) were significantly larger in ME/CFS than GWI and control (p < 0.044 by 2-tailed unpaired t-tests after Bonferroni corrections). Thresholds for significant sensitivities were ≥ 2 out of 20 by receiver operating characteristics indicating that visual and auditory hypersensitivity was common in ME/ CFS and GWI. The sound sensitivity and exaggerated startle responses were consistent with dysfunctional activity in the inferior colliculus [14].

Discussion
The BOLD data are the 2-back>0-back condition that contrasts the difficult high cognitive load continuous 2back working memory task against the simple low cognitive load 0-back stimulus matching attention task [14]. The postexercise and incremental data reflect dynamic effects of exertion on cognition. Exercise caused changes in the 2-back>0-back condition that measures relative brain activation during the more difficult task. Specific effects on 2-back alone, 0-back alone, and 0-back>2-back condition were not assessed here.
The importance of the BOLD data is that there were differences in relative levels of regional blood flow into midbrain nuclei. ME/CFS and GWI had significant incremental changes following exercise whereas controls had no net changes. The changes were opposite in ME/CFS compared to GWI indicating distinct dynamic exercise-induced pathological consequences.
In the original study [14], dorsal midbrain activation was not different between groups prior to exercise. Reanalysis of the same BOLD data using the ascending arousal network nuclei and seed region approach found baseline differences with lower BOLD in ME/CFS than control and GWI ( Figure 1, Table 2). Control had greater activation compared to ME/CFS in VTA, bilateral PTN and L PBC, while GWI was higher than ME/CFS for left PTN and PO.
The pedunculotegmental nuclei (PTN) had reduced activation in ME/CFS. This cholinergic nucleus has extensive efferents that release acetylcholine throughout the cerebrum to maintain wakefulness and induce attention. PTN assists in updating rapidly changing environmental information as required for our continuous 2-back task. Other functions and potential roles in disease dysfunction are discussed in the accompanying paper in this issue.
Exercise caused a significant dynamic switch in midbrain activation. Exercise caused an increase in BOLD in ME/CFS but decrease in GWI. Therefore, exercise had differential effects in ME/CFS compared to GWI. Following exercise, the control was greater than GWI in VTA, bilateral MRF and right PTN, while ME/CFS was higher than GWI for midline PAG, DR and MR, and right lateral MRF, PBC and LC ( Figure 2, Table 3). The seed region approach was consistent with the activation found in the original 141 voxel region of interest that also included inferior colliculus ( Figure S1) [14]. All 141 voxels in the region of interest analysis were contiguous, but the seed region approach assayed the net activation within the smaller volumes defined by the seeds.

Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 23 November 2021
After exercise, the midline nuclei and MRF had significantly lower BOLD activation in GWI than the other two groups.
PAG is integral to threat assessment and instantaneous responses. Detection of a threat activates the PAG and midbrain reticular formation and causes a transition from relaxed wakefulness to high general attention [46]. Active responses range from freeze with conscious tonic immobility if motion would lead to detection by a nearby predator; defensive approach to assess an ambiguous threat (modeled as rumination) [47,48]; flight via an escape route or to shelter; and defensive attack if the predator is within a dangerous distance and escape is not possible [49,50].
VTA is a dopaminergic nucleus that stimulates the locus coeruleus to promote wakefulness [51,52]. It has a role in reward situations and positive emotion. Dysfunction is associated with anhedonia.
DR and MR are serotoninergic nuclei that project to the limbic system during active stress, conflicts and anxiety [53,54]. They may initiate fight or flight decisions. MR participates in tolerance and coping strategies with aversive stimuli as well as arousal, wakefulness and long term memory.
In humans, the MRF is activated during the transition from a relaxed awake state to an attention-demanding state during reaction-time tasks [46] and during focused interrogation of threats while interpreting the proximity of danger (e.g. freezing in place) [55]. The caudal portion of the MRF extends into the cuneiform region that is correlated with cardiovascular dysfunction in ME/CFS [56,57].
It is the predominant source of noradrenergic innervation in the brain [58,59]. Integrated PAG, amygdala and sensory information activate the LC to generate diffuse efferent outputs to the cerebrum and brainstem [60]. They act in an instantaneous fashion like a tripwire for immediate instinctual responses such as freeze-fight-flight, focused cerebral attention and sympathetic activation for immediate action. Inappropriate or dysfunctional activation contributes to anxiety and PTSD [61]. Atrophy of the right locus coeruleus was found at autopsy in veterans with PTSD [62].
Exercise caused no incremental changes in controls (∆BOLD). However, exertion led to significant positive increments in ME/CFS and, by contrast, decreases in GWI. The dynamic changes were significantly different between ME/CFS and GWI for midline PAG, DR, MR and VTA, right LC, and bilateral PTN, PBC, and PO ( Figure 3, Table 4).
The opposite directions of change indicated that distinctly different pathological mechanisms that regulate midbrain blood flow and neurovascular coupling were modulated by exercise in the two diseases indicating that ME/CFS and GWI were excellent "illness controls" for each other.
Both ME/CFS and GWI had significant light and sound sensitivity ( Figure 4). These sensations are monitored in the superior and inferior colliculus, respectively. The inferior colliculi are innervated by the ascending hindbrain auditory pathway (the lateral lemniscus), somatosensory pathways from the medulla, pons, and arousal nuclei [63]. The inferior colliculus participates in multimodal sensory perceptions, vestibulo-ocular reflex, predator aversion and escape, prey localization, social communication, analgesia and fear related behaviours. Sharp acoustic stimulation initiate the startle response with acutely accentuated attention and surveillance leading to visual and truncal orientation towards the sound, generalized hyperarousal and aversive behaviors [64][65][66][67]. Inferior colliculus is relevant to the light and sound sensitivity and heightened startle response in ME/CFS, GWI and veterans with PTSD [68].
The inferior colliculus is highly metabolically active and vulnerable to toxic injury [63]. ME/CFS have reduced cerebral blood flow during heads up tilt [69] and exercise [70] that may reduce oxygen supply to the susceptible inferior colliculus and lead to dorsal midbrain dysfunction. Exercise -induced lability of cerebral blood flow and neurovascular coupling may contribute to the dysfunctional BOLD patterns in the midbrain, insula and cerebellum vermis in GWI and ME/CFS [14] but with different mechanisms and statistically reciprocal outcomes in the two diseases.

Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 23 November 2021
The general linear models assessed the relative influences and interactions between Disease, Orthostatic and gender status on each day and adjusted for age, PTSD, BMI and dolorimetry thresholds. Outcomes for disease status were generally comparable to the ANOVA results.
Orthostatic status was significant before exercise as START and POTS had higher BOLD than STOPP in the left PBC (Table S13). POTS had postural tachycardia before and after exercise, while START were defined by exercise induced postural tachycardia [26,27,31]. After exercise, START had dynamic incremental depression of BOLD activation in the R_LC and bilateral PBC (Table 5) compared to the STOPP subjects who had no changes in BOLD or postural heart rate. Involvement of the locus coeruleus in START implicated exercise-induced autonomic dysfunction. The PBC interrogates pain and interoceptive visceral sensations then forwards the information to the PAG, thalamus, hypothalamus, and amygdala for further processing [71]. PBC is recruited in states of malaise as an adaptive component of the sickness response [72], and so may participate in the experience of postexertional malaise.
Hedges' g ranged from 0.50 to 0.81 which suggests moderate to high effect sizes for replication of the significant results using the same protocols.
It is important to appreciate the limitations of these disease and exercise effects. The findings were inferred based on the seed regions extracted from the ascending arousal network. Coordinates of the seed regions may improve as newer standards are created [40]. The actual metric being compared is the 2-back>0-back differential activity during the difficult cognitive working memory task. The results may not be applicable to the resting state or other cognitive tasks. The neural and vascular responses combined to generate these data without providing insights into functional connectivity or molecular mechanisms. Brainstem motion may blur the borders of nuclei in this seed region approach. Therefore, we consider the results to be a general predictor of changes in BOLD for nuclei in the ascending arousal network that are congruent with the dorsal midbrain region of interest found in our previous study [14]. The results do suggest that significant differences will be found in future studies that specifically target these nuclei in ME/CFS and GWI when suitable sample sizes are compared and advanced motion correction algorithms are applied [73][74][75][76]. The most significant differences were induced by exercise with elevated BOLD in ME/CFS but reductions in GWI, and were most clearly exposed by comparison of ME/CFS versus GWI rather than differences from control subjects (Tables 5 and S19).

Conclusion
The seed region approach based on the ascending arousal network extended our previous finding of exercise induced changes in BOLD during a high cognitive load 2-back working memory task. The salient findings were significantly lower BOLD in the midbrain at baseline in ME/CFS compared to GWI and control, and significant dynamic changes after exercise with elevation of BOLD in ME/CFS but reduction in GWI. Review of the functions of midbrain nuclei provides a fresh perspective on potential neural pathologies affecting inferior colliculus (startle), oculomotor and visual systems, PAG, MRF and other nuclei for threat assessment, anxiety, negative emotion, pain and tenderness and other aspects of the ME/CFS and GWI clinical experiences. The data provide an initial framework to power future studies of postexertional malaise and midbrain dysfunction.

Conflict of Interest
Preprints (www.preprints.org) | NOT PEER-REVIEWED | Posted: 23 November 2021 The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Funding agencies had no influence on the research or manuscript.

Author Contributions
JNB obtained funding, conducted the clinical study, supervised data collection and analysis and prepared the manuscript. AA and HP performed statistical analysis. SDW analyzed fMRI data. All authors contributed to the final version.

Funding
We have no direct or indirect financial, institutional, or personal conflicts of interest to report.