Consecutive patients with ME who visited the Miwa Naika Clinic from September 2018 to May 2022 and were able to stand and walk were included in the study. ME was diagnosed according to the International Consensus Criteria [2].

Patients with significant complications unrelated to ME were excluded from the study, as were pregnant or lactating females. Of the 160 patients, 53 were male and 107 were females, with a mean age of 37 $\pm$ 12 years (range, 18–74 years). The mean length of disease history was 4.8 $\pm$ 6.3 years. Ongoing medications, including nutritional supplements and multienzyme tablets, were continued throughout the study period.

All study participants provided written informed consent. The study was approved by the Toyama Prefectural Medical Association Ethics Committee (approval#: 2016-010) and the Ethics Committee of the Toyama Prefectural Rehabilitation Hospital & Support Center for Children (approval# 47), and was conducted in accordance with the Declaration of Helsinki. Additionally, each participant provided written informed consent for the possible publication of identifying information/images in an online open-access publication.

All study patients underwent neurologic testing, including the Romberg test and tandem gait test to evaluate possible static and kinetic disequilibrium, performance status (PS) scoring, and a conventional active 10-min standing test. Patients with disequilibrium underwent rTMS and were evaluated after treatment.

Participants were evaluated to determine the presence of widespread pain and tenderness, which are characteristic of fibromyalgia [16].

The severity of patient fatigue was scored using the Chalder fatigue questionnaire (CFQ) [17].

Information concerning activities of daily living was obtained from each patient and verified by a physician. PS was graded on a 10-point scale according to symptom severity, as previously reported [4, 5] (Table 1).

To diagnose static disequilibrium, participants were asked to stand with their feet together and eyes closed for 10 s [3, 4, 5]. Unstable standing with wide body sway or oscillations and possibly a fall with feet together and eyes open and/or closed was diagnosed as positive for static disequilibrium [3, 4, 5]. Kinetic disequilibrium was diagnosed using the 2-m tandem gait test. The patient carefully walked 2 m on a straight, flat floor, completing toe-to-heel touch with each step. The patient then turned around and repeated the tandem gait along the same line. Step-out, misstep, and/or body sway with the whole head displaced from the vertical line of gravity at the midpoint between both feet were deemed positive for kinetic disequilibrium and static disequilibrium. None of the 30 healthy control subjects were diagnosed with disequilibrium by the diagnostic criteria used in this study.

The conventional active 10-min standing test was performed as previously reported [5]. Participants were asked to remain standing still with their feet approximately shoulder width apart. Patients were scored positive for orthostatic intolerance when they experienced difficulty during the test due to symptoms such as palpitations, light-headedness, pallor, fatigue, dizziness, and nausea. Because of these symptoms, they were unable to maintain the standing posture and gave up standing.

Patients with both static and kinetic disequilibria and who provided written informed consent were assigned to the rTMS study group. rTMS was performed as described in a previous report [15].

High-frequency rTMS was delivered using a MagstimRapid 2 instrument (Miyuki Giken, Tokyo, Japan) equipped with a figure-of-8 stimulating coil. All rTMS patterns consisted of bursts containing three pulses at 50 Hz and an intensity of 80% of the active motor threshold, repeated at 200-ms intervals (5 Hz). A 2-s train of theta burst stimulation (TBS) pattern was repeated every 10 s for a total of 190 s (600 pulses) in the intermittent TBS (iTBS) pattern [18]. Burst Stimulator (Medical Try System, Tokyo, Japan) is a magnetic stimulation control software employed for the delivery of TBS. The dorsolateral prefrontal cortex (DLPFC) and primary motor area (M1) of the left hemisphere were selected as target areas for iTBS, irrespective of the patients’ right- or left-handedness. DLPFC was identified using magnetic resonance imaging. M1 was determined by measuring the motor threshold. The stimulating coil was first placed on the left DLPFC and then on M1. The stimulation intensity was initially set to 80% of the resting motor threshold, and subsequently reduced according to the patient’s tolerance. The patients underwent rTMS treatment, including 10 sessions each for the DLPFC and M1, over 2 weeks of hospitalization.

Continuous variables are presented as the mean $\pm$ standard deviation. Student’s t-test was used to compare continuous variables. Proportional data were analyzed using Fisher’s exact test with Yates’ correction, and the Mann-Whitney U test was used to compare the median PS scores between groups. Statistical significance was set at a p-value of 0.05.

Among the 160 participants, unstable standing with feet together and eyes open was noted in 7 (4%) cases, and with eyes closed in an additional 33 (21%) cases. Thus, static disequilibrium as demonstrated by unstable standing during the Romberg test was positive in 40 (25%) patients. The duration of ME disease was significantly longer in patients with static disequilibrium than in those without (Table 2). These patients also showed a significantly higher mean CFQ score. Typical Romberg test results for patients with and without static disequilibrium are shown in Fig. 1a–d.

Truncal instability and/or missteps were observed in 57 (36%) of the 160 participants during the first straight tandem gait for 2 m. Typical results for the first straight tandem gait in patients with and without kinetic disequilibrium are shown in Fig. 2a,b. Additionally, instability or missteps in tandem gait after turning 180° was detected in 14 (9%) patients who had normal tandem gait before turning. Fig. 3 shows a typical unstable tandem gait after turning 180° in a patient with kinetic disequilibrium. A total of 71 (44%) patients were diagnosed as positive for kinetic disequilibrium, with a significantly higher prevalence amongst female patients (Table 3). The mean CFQ score was significantly higher in patients with kinetic disequilibrium than in those without.

Medications before the study are shown in Supplementary Table 1. Those medications appeared to be not different between patients with and without static or disequilibrium.

Of the 160 participants with ME, 42 (26%) were diagnosed with fibromyalgia (Table 2). This condition was significantly more prevalent in patients with static disequilibrium (Table 2) and in patients with kinetic disequilibrium (Table 2) than in those without.

The distribution of PS scores in patients with and without static disequilibrium is shown in Fig. 4A, and with and without kinetic disequilibrium in Fig. 4B. The median PS score was significantly higher in patients with static or kinetic disequilibrium than in those without. Moreover, the median PS score was significantly higher in the 39 patients with both static and kinetic disequilibrium than in those without (score of 7 vs. 5, p 0.01).

Orthostatic intolerance was found in 27 (17%) of the 160 participants, as demonstrated by failure to complete the 10-min standing test. Furthermore, orthostatic intolerance was significantly more prevalent in patients with static disequilibrium (55%) than in those without (4%) (Table 2), and in patients with kinetic disequilibrium (38%) compared to those without (0%) (Table 3).

Both static and kinetic disequilibrium were observed in almost all patients with static disequilibrium (39/40, 98%), and in about half of patients with kinetic disequilibrium (39/71, 55%). Only one patient with static disequilibrium exhibited no kinetic disequilibrium. Female prevalence, prevalence of combination with fibromyalgia, prevalence of orthostatic intolerance, mean CFQ score, and median PS score were all significantly higher in the 72 patients with any type of disequilibrium than in the 88 patients showing no disequilibrium (Table 4).

For most patients, the disbalancing motion was not sudden but instead rather slow, as if pulling out, and in one direction only.

The rTMS study was carried out on 13 patients with both static and kinetic disequilibrium. The clinical features of these patients and the therapeutic effects are summarized in Table 4. In most patients, the intensity of stimulation was lowered due to intolerance to discomfort, headache, and dizziness. The CFQ score decreased significantly following treatment (28 $\pm$ 3 to 22 $\pm$ 5, difference: –6 $\pm$ 3, p 0.01). Eleven (85%) patients showed recovery from both static and kinetic disequilibrium after treatment, with a decrease in their PS score of at least 2. Furthermore, 12 patients completed the 10-min standing test after rTMS treatment, whereas only 4 patients completed the test before rTMS. Eight (89%, p 0.01) of the 9 patients with orthostatic intolerance showed recovery after treatment. No patients experienced adverse effects due to the rTMS procedure. Fig. 5 shows typical Romberg and tandem gait test results before and after rTMS treatment in a 30-year-old patient (patient #4 in Table 5). This patient showed both static and kinetic disequilibrium prior to treatment, but became free of any disequilibrium after treatment.

Disequilibrium of central origin is a typical neural sign and includes both static and kinetic disequilibrium. Its prevalence in ME patients is suggestive of a neural disease.

Static and kinetic balance for postural stability is essential for performing daily activities [11, 19, 20]. Contradictory results have been reported with regard to disequilibrium in patients with CFS. Komaroff [21] estimated that 30–50% of patients with CFS experience some degree of disequilibrium, whereas Merry [22] suggested the incidence could be as high as 70% in all patients. Our recent studies suggested that disequilibrium may be involved in the development of orthostatic intolerance [3, 4, 5]. In contrast, Paul et al. [23] found no significant difference in postural stability between patients with CFS and control subjects, as measured by postural sway under eyes-open and eyes-closed conditions. In the present study, standard neurological examinations revealed abnormal results, such as instability when standing with both feet together and eyes open and/or closed, and unstable tandem gait. This suggests disequilibrium or truncal ataxia in a considerable number of patients with ME. A positive Romberg test was previously reported in 10–20% of CFS patients, and an abnormal tandem gait test in 15–25% [6, 7]. The present study found a higher incidence of positive Romberg test (21%) in patients with ME, as well as an abnormal tandem gait test even before turning around (36%). The International Consensus Criteria allows the identification of ME patients who are more physically debilitated and have greater physical and cognitive impairments, thus differentiating patients with ME from those who are depressed [2]. The higher incidence of both a positive Romberg test and an abnormal tandem gait test observed in ME patients may be due to the greater severity of the disease compared to the CFS cases evaluated in previous studies. In the present study, patients with any disequilibrium had higher CFQ and PS scores, suggesting more severe fatigue and restricted activities of daily living. Disequilibrium, or postural instability, is a useful indicator of advanced disease. Recently, Hayes et al. [24] evaluated postural sway during a 30-s static standing test and suggested that people with long coronavirus disease (Long COVID) and those with ME/CFS have similarly impaired balance and physical capacity.

Static balance and postural stability are regulated by a postural adjustment system [19, 20]. Multiple types of sensory information, such as somatosensory, visual, and vestibular sensations, are processed by various brain areas in the multisensory cortical area and then in the corticocortical vestibular network to achieve adaptable postural control. The automatic process of postural control is termed postural reflex and includes head-eye coordination accompanied by appropriate body segment alignment [19, 20]. This control is mediated by descending pathways from the brainstem, with cooperation between the vestibulospinal, reticulospinal, and tectospinal tracts contributing to the process. The visual sensory-mediated system can compensate for central vestibular function to maintain postural control when the above-mentioned system is impaired in the brain cortex. Visual information from the eyes is projected directly to the vestibular and oculomotor nuclei in the midbrain to stimulate the head and external oculomotor muscles and maintain postural stability, thereby helping vestibular function in cooperation with the cerebellum. Visual compensation is therefore likely to be essential for maintaining postural stability in the presence of vestibular deficits [19, 20], explaining why many ME patients lost balance during the Romberg test when their eyes were closed. Regulation of kinetic stability is usually more difficult under locomotive conditions than stability under static conditions, especially with tandem gait and turning 180°, even with eyes open.

The exact cause of the disequilibrium observed in patients with ME remains unknown. However, it appears to have a central vestibular origin, consistent with previous results of vestibular function tests in patients with CFS [25, 26]. The pathogenesis for the observed neurologic defects in disequilibrium may be due to neural brain inflammation [27].

Previous studies have demonstrated significantly impaired functional connectivity or default mode network functional hypo-connectivity in the brain of ME/CFS patients [28, 29, 30]. Furthermore, the symptom severity of such patients was associated with functional connectivity in diverse brain regions. It was recently reported that COVID-19 infection may have a widespread effect on the functional connectome. This is characterized by lower functional connectivity and altered patterns of information processing efficiency and effective information flow, similar to ME/CFS [31].

The long-lasting neuromodulatory effects of rTMS are of great interest for therapeutic applications in various neurological and psychiatric disorders where the functional connectivity among brain regions is profoundly disturbed [32, 33, 34, 35]. rTMS and TBS show significant promise in improving the outcomes for adults with complex neurological and neurodegenerative conditions, such as Alzheimer’s disease, stroke, and Parkinson’s disease [35]. rTMS and TBS may impact neuroplasticity and functional connectivity, in particular through brain derived neurotrophic factor and tropomyosin-related kinase receptor type B. The plastic after-effects of rTMS result in large-scale changes at the network level that affect both local and remote activity within the stimulated network, as well as interactions between the stimulated and distinct functional networks.

The cortical organization of human vestibular information processing is poorly understood. rTMS of the DLPFC and M1 may facilitate the recovery of central vestibular function by enhancing neural network activity within the known vestibular function center sites located in the brainstem and in the cortex. The vestibular system provides information on head translation, rotation, and orientation in a gravitational environment [36]. The corticovestibular network is distributed throughout the brain and has a high degree of functional connectivity that can quickly buffer the network and reestablish network integrity in case of injury [36, 37, 38]. Central or cortical vestibular region disorders or lesions, such as cerebrovascular events and encephalitis, rarely cause vestibular deficits in the acute phase. Moreover, the symptoms do not persist, probably because of a highly robust and redundant vestibular cortical network. Neural inflammation in ME encompasses almost the entire central nervous system [37]. This is probably why disequilibrium associated with central vestibular dysfunction develops and persists in many patients with ME. The DLPFC and/or M1 appear to be extremely important modulators of this neural network. rTMS treatment probably corrects or enhances neural network function through the vestibular center, resulting in the amelioration of disequilibrium.

Kakuda et al. [39] conducted a pilot study in several patients with CFS in which they applied high-frequency rTMS over the DLPFC. They observed favorable therapeutic effects and reduced fatigue symptoms in most cases, and speculated the beneficial effect of rTMS was mediated through DLPFC neural activation. Hence, recovery or enhancement of the impaired corticocortical network may play a crucial role in the therapeutic effects of rTMS.

The present study found a close relationship between disequilibrium and orthostatic intolerance, as suggested by other recent reports. However, cerebral hypoperfusion and exaggerated sympathetic activation are thought to be the primary pathogenesis of orthostatic intolerance. A paradigm shift is therefore required [5].

Kinetic disequilibrium is also associated with gait disturbances. Patients with kinetic disequilibrium may experience difficulty walking in crowded streets, and need to make small, frequent and rapid changes in walking direction, as well as frequent and rapid changes in walking speed. Such patients are unable to walk for extended periods of time because of the effort required, and hence they soon become fatigued.

Fibromyalgia is significantly more prevalent in patients with static or kinetic disequilibrium than in those without. A recent study in patients with CFS suggested that poor balance was associated with poor self-reported physical health [40]. Although none of the patients in the present study reported that their disequilibrium was due to musculoskeletal pain, I cannot exclude that pain may have influenced the neurological tests results. Differences between the pathophysiology of ME in combination with fibromyalgia, and ME alone, remain to be elucidated and require further investigation.

This study had several limitations. First, there were no control patients with sham treatment for comparison with rTMS treatment. Second, rTMS treatment was applied to both the DLPFC and M1 in all participants. Consequently, the effect of this treatment could not be evaluated individually for the DLPFC and M1. Third, the long-term effects of rTMS treatment were not determined.