Transcranial magnetic stimulation uses dynamic magnetic fields to generate induced currents to regulate individual neurons and neuron groups in the cortex and the neural networks connected to them (Young et al., 2014). Strong effects can sufficiently depolarize neurons to trigger action potentials (Young et al., 2014). If TMS pulse stimulation is repeatedly given, it is called repetitive TMS (rTMS).

The pathophysiological mechanism of NP may be related to changes in the structural or functional plasticity of the central nervous system (Leung et al., 2009). The maintenance of NP mainly depends on central sensitization. Central sensitization refers to the abnormal increase in excitability or synaptic transmission of central pain-related neurons, including the increase in spontaneous discharge activity of neurons, the expansion of sensory domains, and the reduction of the threshold value to external stimuli, thus amplifying the transmission of pain signals (Nickel et al., 2012). rTMS acts on the cerebral cortex and adjacent structures under the cortex through high-level regulation of the central nervous system, and it exerts various effects on the pain process, thereby exerting an analgesic effect. Yang et al. (2018) suggested that HF-rTMS may reduce central sensitization and relieve NP by downregulating the overexpression of neuronal nitric oxide synthase in ipsilateral dorsal root ganglions and inhibiting the activity and proliferation of astrocytes in the L4–L6 spinal dorsal horns ipsilateral to NP.

A meta-analysis of 25 studies (589 long-term follow-up patients) evaluated the efficacy of HF-rTMS for the treatment of NP (Jin et al., 2015). Pooled analgesic results showed a statistically significant effect size of −0.86 (p < 0.05), indicating that rTMS can effectively reduce the pain intensity of NP from different sources. The result indicated that a single rTMS treatment can remarkably reduce the pain intensity of patients with NP. When the number of sessions was increased from 2 to 10, the subjects also produced considerable pain relief, especially among patients with pain after suffering from central stroke. After 1–2 months of follow-up (161 participants), the analgesic effects of multiple rTMS treatments (≥5 sessions) were observed to last at least 1 month but not more than 2 months, indicating that the different analgesic effects of rTMS may depend on the neuroanatomical source of the pathophysia of NP, that is, the more effective source of the therapeutic effects of rTMS on NP is the “top” (supraspinal, cranial, or spinal) rather than the “bottom” (nerve roots or peripheral nerves). Similarly, Zhang et al. (2021) systematically reviewed 29 studies (24 for rTMS, 736 participants), and found that rTMS successfully improved the pain symptoms of 97.1% of patients with NP (715 participants). The analgesic effect of rTMS was maintained for 2 weeks after the last session, but this beneficial effect usually lasted for less than 1 month. In addition, the M1 region was targeted in almost all patients (82.5%, 607 participants).

As recent reviews have concluded, the application of HF-rTMS of M1 may alleviate various types of NP. Some of the NP conditions with the greatest response to rTMS include post-stroke central pain and trigeminal neuralgia, whereas NP conditions with more peripheral anatomical origins, such as post-traumatic peripheral NP, are less reactive to rTMS (Lefaucheur, 2016; Leung et al., 2020). Although a few studies suggest this conclusion, we think it is premature to present this type of assertion which is not based on published data on large series. It is unclear whether any particular type of NP would be considered a better indication for rTMS treatment.

Fibromyalgia is usually insensitive to conventional treatment. Previous studies indicated that rTMS may work by modulating pain pathways (Bestmann et al., 2004; Yoo et al., 2008), such as the descending inhibitory pathway, and by modulating social–affective areas of the brain, such as the right temporal lobe (Boyer et al., 2014). A meta-analysis (7 studies, 210 patients) evaluated the efficacy of HF-rTMS of M1 in the treatment of fibromyalgia (Saltychev and Laimi, 2017). Based on the patients’ scores on a 0–10 numerical rating scale, their pain intensity before and after the last rTMS session decreased by 1.2 points. Moreover, the pain intensity before the last stimulation and 1 week to 1 month after the last stimulation decreased by 0.7 points. Both pooled results were statistically significant but below the cut-off point of a minimum clinically significant difference of 1.5 points. In a narrative review that analyzed 12 studies on fibromyalgia, 9 studies concluded that rTMS of M1 was effective in relieving pain in patients with this condition (Yang and Chang, 2020). By contrast, the remaining three randomized controlled trials (RCTs) denied that patients with fibromyalgia benefited from rTMS.

Although the results of some studies were negative, the fact that fibromyalgia is difficult to manage suggests that rTMS is a potential analgesic method for managing fibromyalgia.

Recent studies suggested that the mechanism of migraine may be linked to neurological causes, such as cerebral cell hyperexcitability and altered cortical excitability (Cosentino et al., 2014; Lan et al., 2017). On the one hand, rTMS can regulate the excitability of cortical structures involved in pain control, such as inhibiting cortical spreading depression (Chervyakov et al., 2015). Therefore, rTMS may contribute to the prevention and relief of headache symptoms during migraine attacks. On the other hand, rTMS can promote the release of endogenous analgesic substances, such as dopamine and endogenous opioid peptides, thereby relieving headaches.

Currently, in the United States, acute migraine with aura is the only FDA-approved indication for single-pulse TMS45. A meta-analysis also reported that single-pulse TMS was effective in the acute treatment of aura migraine after the first attack (p = 0.02), but not in chronic migraine (p = 0.14) (Lan et al., 2017). However, another systematic review found that TMS and rTMS contributed to reductions in headache frequency, duration, intensity, abortive medication use, depression, and dysfunction in both chronic primary and secondary headaches (Stilling et al., 2019). In addition, consistent with the findings of Lan et al. (2017), only a few studies reported greater changes than sham stimulation (Stilling et al., 2019). Similarly, another systematic review found that, compared with sham stimulation, the outcome indicators of patients suffering from migraine who received HF-rTMS treatment substantially improved, including headache frequency, pain intensity, headache duration, and dosage (Yang and Chang, 2020). However, two studies showed that the results after rTMS were not superior to those after sham stimulation, and both showed a strong placebo response. Despite conflicting evidence on its efficacy, rTMS may still be a potential option for patients with migraine.

Previous studies have demonstrated that abnormal postural control of trunk muscles may lead to the occurrence of chronic low back pain (CLBP) (Deliagina et al., 2006). Furthermore, the M1 region is believed to play a key role in postural control regulation (Ambriz-Tututi et al., 2016). Ambriz-Tututi et al. (2016) evaluated the long-term effects of rTMS on CLBP and compared them with those of physical therapy and sham stimulation. Results showed that the patients who received HF-rTMS of M1 had a remarkable decrease in pain intensity after 1 week of treatment. After 3 weeks of continuous treatment, their pain intensity was reduced by nearly 80% compared with the baseline and was substantially lower than that of the group that received sham stimulation or physical therapy. In addition, a previous study reported that, compared with sham stimulation, 1 session of HF-rTMS can significantly lessen the pain intensity felt by patients with CLBP (p < 0.001) (Johnson et al., 2006). Compared with those on fibromyalgia and headache, there is less evidence to suggest the efficacy of rTMS on CLBP.

In Canada, Australia, Japan, the European Union, and Israel, TMS devices have been approved for depression, schizophrenia, and NP (Marangell et al., 2007). However, in the United States, the application of TMS to pain management is only considered investigational.

Thus far, most high-quality RCTs and systematic reviews have shown that rTMS applied to the dorsolateral prefrontal cortex (DLPFC), the supplementary motor area (SMA), or the primary somatosensory cortex (S1) lacks analgesic effects, while stimulation of M1 provides pain relief (Marangell et al., 2007; Lefaucheur et al., 2020). In 2020, the guidelines for the use of rTMS developed by an expert panel in Europe stated that HF-rTMS of M1 has a clear effect on NP and fibromyalgia (level A evidence) (Lefaucheur et al., 2020). The selection of stimulation targets and stimulation parameters is the most basic and important issue that determines the analgesic effect of rTMS. Some of the current issues in stimulating targets are as follows: (1) whether different types of CP conditions require specific stimulation targets; (2) whether combined treatment with different stimulation targets will enhance their analgesic effect; (3) lack of evidence from studies that compared the analgesic effects of different stimulation targets; and (4) few use image-guided navigation to improve the accuracy and repeatability of targeting. Another potential concern with rTMS is the duration effect. The number of sessions in most TMS studies ranged from 1 to 10 sessions. Some studies have reported that the analgesic effects of rTMS are cumulative and require multiple sessions to achieve clinically significant effects (Brighina et al., 2009). A single rTMS session may not be sufficient to induce changes in cortical excitability, and multiple sessions are required to induce changes in neuroplasticity. However, prolonged use of rTMS to increase stimulation intensity may also lead to a reduction or even reversal of the stimulatory effect of motor cortical excitability, making the treatment counterproductive. Therefore, the number of sessions with rTMS should be investigated to provide best practice for the analgesic effects of rTMS.

Transcranial direct current stimulation works by using two or more electrodes to apply a low-amplitude direct current (typically from 0.5 to 2 mA) to specific brain regions to modulate their excitability, thereby relieving pain (Nitsche and Paulus, 2000). Traditional tDCS usually uses two sponge electrodes, one as the anode and the other as the cathode. HD tDCS utilizes a set of smaller electrodes, such as 4 × 1, to provide more focused stimulation.

In addition to network effects and changes in central nervous excitability, tDCS may also alleviate NP by regulating the central nervous immune system and inhibiting glial cell activation. Cioato et al. (2016) investigated the effects of tDCS on nociceptive responses and measured IL-1 β, IL-10, and TNF-α levels in the central nervous system structure of NP rats. After tDCS, the levels of IL-1β and TNF-α decreased, whereas those of IL-10 increased, suggesting that tDCS may modulate the immune system to alleviate NP. Recent studies have established that microglia and astrocytes in the nervous system play key roles in the initiation and maintenance of NP, respectively (Gritsch et al., 2016). As a common method for regulating cortical excitability in superficial pain-related areas, tDCS can inhibit neuronal sensitivity after peripheral nerve injury and downregulate the expression of the P2 × 4 receptor, thereby inhibiting microglia activity, ultimately leading to NP remission (Zhang et al., 2020).

A systematic review (8 studies, 127 participants) suggested that, compared with sham stimulation, tDCS could notably reduce pain intensity in patients with NP associated with spinal cord injury (SCI), stroke, and amputation, and its analgesic effect lasted for 1 week after the end of the intervention (David et al., 2018). However, no significant differences between the groups were observed in patients with radiculopathy. Similarly, a meta-analysis reported a moderate effect of tDCS in reducing NP in patients with SCI; however, the effect was not maintained at follow-up (Mehta et al., 2015). A mean pooled decrease of 1.33 units on a 10-item scale was found post treatment. In another systematic review (6 studies, 125 patients with NP), 5 studies found that, compared with sham stimulation, anodal tDCS remarkably alleviated NP (Zhang et al., 2021). Overall, many studies have confirmed that tDCS has a moderate effect on pain relief among individuals with chronic NP, but this effect is not maintained during follow-up.

The pathogenesis of fibromyalgia may be related to the dysfunction of the central nervous system (Cook et al., 2007; Brown et al., 2014). As a neuromodulation technique that targets the central nervous system, tDCS may theoretically help relieve pain. A meta-analysis of 6 studies (192 patients with fibromyalgia) reported that, compared with sham stimulation, anodal tDCS of M1 was more likely to relieve pain and improve fibromyalgia-related function, whereas cathodal tDCS of M1 and anodal tDCS of left DLPFC did not produce notable analgesic effects (Zhu et al., 2017). Although this meta-analysis was unable to calculate the overall effect of tDCS on fibromyalgia during the follow-up period, some of the studies it included concluded that 10 sessions of anodal tDCS over M1 was more likely to control pain than 5 sessions, and this effect might last up to 2 months. Similarly, Lloyd et al. (2020) also reported that active tDCS applied at an intensity of 2 mA to left M1 for 20 min/days for 10 sessions appears to be able to lower pain intensity in fibromyalgia. Hou et al. (2016) reviewed 16 studies (5 for tDCS, 11 for rTMS; 572 patients with fibromyalgia), found that aside from improving cognitive function, both tDCS and rTMS had similar positive effects on pain symptoms, sleep disturbances, and tender spots. However, rTMS produced a greater analgesic effect than tDCS. Therefore, excitatory rTMS/tDCS should be considered in the treatment of patients with fibromyalgia, especially for those with painful symptoms that are not responding to other therapies or for whom the continuation of such therapies is not possible due to their adverse side effects (as is commonly the case with FDA-approved drugs).

A systematic review of 12 studies on chronic headache (8 for migraine, 413 participants) found that, compared with baseline, tDCS substantially reduced headache frequency in 7 studies (Stilling et al., 2019). However, only one study showed that, compared with sham stimulation, tDCS considerably decreased headache frequency. Six studies reported that tDCS shortened headache duration, but only one study was statistically different from the control group. Seven studies found that tDCS reduced pain intensity, but only two studies showed significant differences between groups. Similarly, Feng et al. (2019) reviewed 9 studies (4 for tDCS, 115 participants), and found that anodal tDCS of M1 markedly reduced the frequency and intensity of headaches in patients suffering from migraine. Moreover, tDCS over DLPFC substantially reduced pain intensity in these patients, but it had no notable effect on attack frequency. In addition, compared with sham stimulation, cathodal tDCS applied to the vertex or visual cortex did not remarkably change the frequency and intensity of headaches in these patients. Compared with those on fibromyalgia and NP, there is less evidence to suggest the efficacy of tDCS on migraine.

Unlike acute low back pain, non-specific CLBP usually has no peripheral cause (Latremoliere and Woolf, 2009). Central mechanisms have been hypothesized to explain the development and maintenance of pain. Great functional connectivity between the dorsal medial PFC–amygdala–accumbens circuit in patients with subacute low back pain contributes to the risk of CP (Vachon-Presseau et al., 2016). Therefore, brain network disturbance is considered one of the possible causes of CLBP. A recent systematic analysis (eight studies) revealed that, compared with sham stimulation, 1 session of tDCS treatment resulted in substantial pain relief (Patricio et al., 2021). By contrast, multiple sessions of tDCS treatment did not improve short-term and medium-term pain. In 2019, based on two studies on tDCS, an expert panel proposed a level A recommendation against the use of tDCS of M1 for the treatment of CLBP (Baptista et al., 2019).

Recent systematic analyses did not support the use of tDCS for CLBP treatment, and evidence of low quality suggests that tDCS negatively affects CLBP. A possible explanation for these negative findings is that the pathogenesis of CLBP is affected by multiple factors. Therefore, the participants enrolled in these reviews might have had other mechanical diseases or complications, such as cervical spondylosis or small joint disorders. In addition, visceral pain involving the lower back can be a misleading condition, leading doctors to misdiagnose or miss a diagnosis.

In the United States, tDCS has not been approved for any clinical indications but is only used as a research technique for pain management. The guidelines for neurostimulation therapy of CP issued by the European Academy of Neurology gave “weak recommendations” for the use of tDCS for the treatment of peripheral NP and “uncertain recommendations” for the treatment of fibromyalgia (Cruccu et al., 2016; Knotkova et al., 2021). In the United States, tDCS has not been approved for any clinical indications but is only used as a research technique for pain management. The effectiveness of tDCS in relieving CP may vary according to pain subtype, including spontaneous, paroxysmal, and persistent pain (Soler et al., 2010). Thus far, many high-quality RCTs and systematic reviews have shown that the application of tDCS to DLPFC, SMA, or S1 lacks analgesic effects, whereas stimulation of the M1 region provides pain relief (Lefaucheur et al., 2017). It is worth noting that tDCS is non-local with a network effect, and cortices adjacent to the stimulation target may also be affected. As a result, it is essential to combine tDCS with neuroimaging and functional connectivity analysis so that we can attribute specific efficacy to neuromodulation of M1 only, but unfortunately, the majority of tDCS studies lack the corresponding neuroimaging evaluation. In addition, the current level of evidence supporting the positive effect of the application of tDCS over M1 on pain relief is considerably lower than that of rTMS. Compared with that of rTMS, the after-effect of tDCS is also less obvious (Ngernyam et al., 2013).