ISSN 0017-8748

Headache doi: 10.1111/head.13645

© 2019 American Headache Society Published by Wiley Periodicals, Inc.

Review Article
Effects of Non-invasive Brain Stimulation on Headache
Intensity and Frequency of Headache Attacks in Patients With
Migraine: A Systematic Review and Meta-Analysis
Yali Feng, BSc*; Bingbing Zhang, MSc*; Jiaqi Zhang, MSc; Ying Yin, PhD

Background.—Non-invasive brain stimulation (NIBS) techniques such as repetitive transcranial magnetic stimulation (rTMS),
as well as transcranial direct current stimulation (tDCS) electrically stimulate the brain and modify brain activity to suppress
pain. This method is emerging as a potential clinical intervention against migraine.
Objective.—To quantitatively review the efficacy of rTMS and tDCS in randomized controlled trials (RCTs) in modifying
headache intensity and frequency of headache attacks in patients suffering from migraine.
Methods.—We searched 5 databases: PubMed, EMBASE, Web of Science, Cochrane Library, and Scopus for articles from
January 2000 to September 2018. Any RCT regarding the efficacy on rTMS and tDCS on patients with migraine, was con-
sidered to be included. A random effects meta-analysis was performed to pool effect sizes of outcomes related to headache
intensity or frequency of headache attacks. The methodological quality of the included RCTs was assessed using the Physiotherapy
Evidence Database scale.
Results.—Nine RCTs with 276 participants in total (experimental group [EG]  =  149; control group [CG]  =  127) were
included in this review. Five included articles used rTMS (EG  =  81; CG  =  80), and 4 used tDCS (EG  =  68; CG  =  47).
Meta-analysis of excitatory primary motor cortex (M1) stimulation showed significant effects on reducing headache intensity
in patients with migraine (Hedges’ g  =  −0.94; 95% CI, −1.28 to −0.59; P  <  .001, I2  =  18.39%) with a large effect size.
Meta-analysis of excitatory M1 stimulation showed significant effects on reducing frequency of headache attacks in patients
with migraine, with a large effect size (Hedges’ g  =  −0.88; 95% CI, −1.38 to −0.38; P  =  .001, I2  =  57.15%). Excitatory
dorsolateral prefrontal cortex stimulation showed a significant effect on the headache intensity in patients with migraine (Hedges’
g  =  −1.14; 95% CI, −2.21 to −0.07; P  =  .04, I2  =  61.86%) with a large effect size. However, reductions of frequency of
headache attacks was not significant.
Limitations.—Potential differential effects of rTMS and tDCS, various sham methods, and potential overlapping headache
disorders among included subjects may affect the estimation of effect sizes.
Conclusion.—Excitatory NIBS of the M1 is likely to reduce headache intensity and the frequency of headache attacks in
patients with migraine.

Key words: non-invasive brain stimulation, transcranial magnetic stimulation, transcranial direct current stimulation,
migraine, pain

Abbreviations: CI confidence interval, cTBS continuous theta burst stimulation, DLPFC dorsolateral prefrontal cortex,
fMRI functional magnetic resonance imaging, iTBS intermittent theta burst stimulation, LTD long-term
depression, LTP long-term potentiation, M1 primary motor cortex, NIBS non-invasive brain stimulation,

From the Department of Rehabilitation Medicine,  The Second Affiliated Hospital of Chongqing Medical University, Chongqing,
China (Y. Feng and Y. Yin); Department of Rehabilitation Medicine,  The Fifth Affiliated Hospital of Zhengzhou University,
Zhengzhou, China (B. Zhang); Department of Biomedical Engineering,  The Hong Kong Polytechnic University, Hong Kong, China
(B. Zhang); Department of Rehabilitation Sciences,  The Hong Kong Polytechnic University, Hong Kong, China (J. Zhang).

Address all correspondence to Y. Yin, Department of Rehabilitation Medicine, The Second Affiliated Hospital of Chongqing Medical
University, Chongqing, China, email: 300735@cqmu.edu.cn

Accepted for publication August 11, 2019.

1436
Headache
PEDro physiotherapy evidence database, PICOS population, intervention, comparison, outcomes, and
study design, RCTs randomized controlled trials, rTMS repetitive transcranial magnetic stimulation, sTMS
single-pulse transcranial magnetic stimulation, tDCS transcranial direct current stimulation
(Headache 2019;59:1436-1447)
INTRODUCTION
Migraine is one of the most prevalent neurologi-
cal disorders that has a significant effects on patient
quality of life.1 Migraine sufferers experience periodic
headache attacks of varying degrees that are associated
with other symptoms such as vomiting, photophobia,
and sound sensitivity.2,3 Behavioral interventions and
medications are the 2 most common management
strategies against migraine, but these strategies are
often limited in efficacy. Poor compliance of migraine
sufferers to behavioral interventions results in limited
success, and medications come with significant and
apparent side effects that has a significant impact on
patient compliance and adherence.4,5 It is therefore
imperative that other alternative therapeutic strategies
for migraine are developed.
Neuromodulation by non-invasive brain stimu-
lation (NIBS) has been proposed as a possible non-
pharmacological intervention for migraine in recent
years.6 Through a phase II randomized controlled
trial (RCT), single-pulse transcranial magnetic stim-
ulation (sTMS) has been identified as an efficacious
method in the acute treatment of migraine associated
with aura. This has been approved by the U.S. Food
and Drug Administration (commonly known as the
FDA) as a treatment for migraine. This technique is
based on the scientific rationale that sTMS can block
cortical spreading depression, also known as spread-
ing depolarization, which is assumed to be the source
of headache in migraine patients.7 Multiple sessions of
repetitive transcranial magnetic stimulation (rTMS)
and transcranial direct current stimulation (tDCS)
have also been used widely in clinical settings as they
are believed to induce long-lasted brain modulatory
effects, for instance by inducing long-term potentiation
*
These authors contributed equally.
Conflict of Interest: None
Funding: This work was supported by National Natural Science
Foundation of China [grant number 81702221].
1437
(LTP) or long-term depression (LTD)-like brain plas-
ticity changes.8,9 The primary difference between these
2 methods is through the delivery mechanism; tDCS
delivers electric currents to the brain tissues through
electrodes placed on the scalp,10 while rTMS uses
repeated magnetic pulses to induce electric currents
in brain tissue through electromagnetic conduction.
Two subtypes exist: anodal tDCS and high-frequency
rTMS facilitates cortical excitability, while on the con-
trary, cathodal tDCS and low-frequency rTMS sup-
presses this cortical excitability.11
NIBS for the management of pain may be
effective in various parts of the cortex.12 For instance,
motor cortex stimulation is widely used for analgesia; a
wealth of evidence has demonstrated that stimulation
of the primary motor cortex (M1) inhibits the activity
of the thalamus, which is related to pain perception.13
One other area that is commonly targeted in NIBS
pain-alleviating strategies is the dorsolateral prefron-
tal cortex (DLPFC). It has been previously shown that
activation of the DLPFC may inhibit the activity of
the midbrain-medial thalamic pathway, which is con-
nected to pain release.14 In patients with migraine,
functional magnetic resonance imaging (fMRI) studies
have shown hypermetabolic activity within the visual
cortex during interictal periods of migraine;15-20 pho-
sphene thresholds of the visual cortex are also found
to be significantly lower in patients suffering migraines
compared to those who do not.17,20 These findings
indicate that the visual cortex is associated with hyperac-
tivity and may also be a possible target for pain allevi-
ating strategies for patients with migraine.
However, a substantial number of clinical trials
concerning the efficacy of NIBS in migraine have been
published with conflicting results. While different brain
stimulation targets have been investigated in migraine
research including M1, DLPFC, and visual cortex,
there lacks a focused review that quantitatively assesses
the effects of different NIBS protocols in the treatment
1438
of migraine. The aim of this systematic review and
meta-analysis is to investigate the effects of NIBS- in
particular rTMS and tDCS- on reducing headache
intensity, and the frequency of headache attacks in
patients with migraine.
METHODS
Literature Search.—A literature search was con-
ducted for studies published from January 1, 2000 to
September 30, 2018 of studies indexed in 5 electronic
databases which include PubMed, EMBASE, Web
of Science, Cochrane Library, and Scopus. The
keywords used for identifying TMS are transcranial
magnetic stimulation, repetitive transcranial magnetic
stimulation, TMS, and rTMS. The keywords used
for identifying tDCS are transcranial direct current
stimulation and tDCS. The keywords used for identi-
fying TBS include continuous theta burst stimu-
lation, intermittent theta burst stimulation, cTBS,
and iTBS. The keywords used for identifying NIBS
include non-invasive brain stimulation and non-invasive
brain  stimulation. The keywords used for identifying
migraine include migraine disorders, migrain*,
headache*, and hemicran*. The reporting in this review
followed the Preferred Reporting Items for Systematic
Review and Meta-Analysis.21 The protocol used in this
review was not registered prior to analysis.
Inclusion and Exclusion Criteria.—We followed the
PICOS framework to organize the inclusion criteria.
Population (P): studies that recruited adult participants
with migraine; Intervention (I): NIBS, including
rTMS and tDCS; Comparison (C): sham stimulation
control or non-interventional control; Outcomes (O):
outcome related to headache intensity or frequency of
headache attacks; and Study design (S): randomized
controlled trials.
Studies meeting any of the criteria were excluded:
(1) they recruited subjects with other headache disor-
ders other than migraine; (2) they were published as
conference abstracts, dissertations, or in books; (3)
studies without sufficient data to estimate the effect
size and the authors did not provide us with data upon
our request; and (4) studies where the participants in
control group were healthy people.
Quality Assessment.—The quality of the included
RCTs was assessed by the Physiotherapy Evidence
October 2019
Database (PEDro) scale, which provides a more
complicated assessment for the quality of studies,
in contrast to Jadad scale.22 A comparison between
PEDro and Jadad scales was presented in our
supplementary section. The PEDro scale includes 10
items, including random allocation, concealment of
allocation, baseline equivalence, blinding procedure
(of subjects, therapists and assessors), adequate follow-
up, intention-to-treat analysis, between-group statistical
analysis, measurement of data variability, and point
estimates. Two authors (YF and BZ) independently
evaluated each article. Any scoring discrepancies were
resolved by a discussion with the third author (JZ).
Data Extraction.—After identifying relevant stud-
ies, 2 authors (YF and BZ) independently extrac-
ted the following information from each article:
(1) the diagnostic criteria of participants; (2) the group
allocation of participants; (3) the stimulation proto-
col; and (4) outcomes of interest and assessment time
points. Any disagreement was settled by a discussion
with the third author (JZ).
Data Analysis.—All statistical analyses were per-
formed using the Comprehensive Meta-Analysis Soft-
ware (version 3.0). The change scores were extracted
in the estimation of effect sizes because of the potential
baseline difference across groups. Hedges’ g was
calculated as the estimation of effect sizes in all meta-
analyses, since headache intensity and frequency of
headache attacks were calculated in different ways across
our included studies.23 A correction of bias to the
pooled variances used in Hedge’s g was applied
because of the small sample size among the included
studies. Authors were contacted by email in the case of
missing or incomplete data. For graphically reported
data, a graph digitizer (GetData Graph Digitizer) was
used to extract the necessary data from the reported
figures. A random effects model meta-analysis was
performed because of the methodological heterogeneity
across studies. Meta-analysis was performed based on
different brain stimulation targets (M1 vs DLPFC
vs visual cortex) and different properties of NIBS
(excitatory vs inhibitory).
Statistical heterogeneity was examined by the I2
statistic. Studies with an I2 of 25-50% were considered
to have low heterogeneity, I2 of values of 50-75%, and
>75% were considered indicative of moderate and high
Headache
Fig. 1.—Flow chart of study selection. [Color figure can be viewed at wileyonlinelibrary.com]
level of heterogeneity, respectively.23 The statistical
threshold was set at P < .05. Where significant results
were found, sensitivity analysis was performed. The
first sensitivity analysis was performed by the leave-
one-out method. Subsequent sensitivity analyses were
performed based on the methodological quality. We
removed studies which were not subject-blinded, asses-
sor blinded, or studies without an intention to treat
analysis for missing cases in turn, to order to see the po-
tential impact of them on the estimation of effect sizes.
Subgroup analysis was performed by only analyzing
the studies in which participants received medication
against migraine during the intervention of NIBS.
RESULTS
Identification and Selection of Studies.—The details
of study selection are shown in Figure 1. Our initial
search yielded 3123 results. After the removal
of duplicated recordings, 1858 citations were iden-
tified and 36 citations remained after reading
1439
titles and abstracts. Twenty-seven articles were
excluded because of the following reasons: 19 studies
were quasi-randomized or non-randomized studies, 3
studies were only study protocols, 2 studies were found
to have used real stimulation in both experimental and
control groups, and 3 studies were found to have
insufficiently reported data for meta-analysis. Further
enquiries to the authors for supplementary data were
unsuccessful. Ultimately, 9 studies were included in
our review.
Characteristics of Included Studies.—The charac-
teristics of included studies are described in Table 1.
There are 5 articles, with a total of 161 participants
where the effects of rTMS on migraineurs were
investigated.15,24-27 Four studies used high-frequency
stimulation15,25-27 and 1 study used low-frequency
stimulation.24 Four articles with a total of 115 partici-
pants explored the effects of tDCS on migraineurs,28-31
of which 3 studies used anodal tDCS28,29,31 and 1 study
used cathodal tDCS.30 Three studies targeted the
 Table 1.—Characteristics of Included Studies
1440

Group Active Feature of Brain Frequency Type of Assessment

Study Participants Allocation Stimulation Stimulator Target Intensity (Hz) Total Dose Control Time Points Main Outcome

Brighina et al15 Chronic migraine EG = 6 rTMS Figure -of- LDLPFC 90% RMT 20 12 sessions Coil rotation Pre Headache frequency
(IHS) CG = 5 eight coil 4800 pulses (vertical to 1-m Headache index
LDLPFC) Post
Teepker et al24 Migraine (IHS) EG = 14 rTMS Round coil Vertex 100% RMT 1 5 sessions Sham coil Pre Headache frequency
CG = 13 5000 pulses Post Pain intensity
(0-10)
Auvichayapat et Migraine EG = 20 Anodal tDCS 35 cm2 LM1 1 mA NA 20 sessions Sham tDCS Pre Headache frequency
al29 (ICHD-2) CG = 17 electrode 400 minutes Post Pain intensity
8-w FU VAS (0-10)
12-w FU
Dasilva et al28 Chronic migraine EG = 8 Anodal tDCS 35 cm2 M1 2 mA NA 10 sessions Sham tDCS Pre Pain intensity
(ICHD-2) CG = 5 electrode 200 minutes 30-d VAS (0-10)
Post
60-d FU
Misra et al25 Migraine (IHS) EG = 47 rTMS Figure -of- LM1 70% RMT 10 3 sessions Sham coil Pre Headache frequency
CG = 48 eight coil 1800 pulses 1-w Pain intensity
2-w VAS (0-3)
3-w
Post
Conforto et al27 Chronic migraine EG = 7 rTMS Figure -of- LDLPFC 110% RMT 10 23 sessions Coil rotation Pre Headache frequency
(ICHD-2) CG = 7 eight coil 36800 pulses (vertical to 4-w Pain intensity
vertex) Post (0-10)
Rocha et al30 Migraine EG = 10 Cathodal 35 cm2 V1 2 mA NA 12 sessions Sham tDCS Pre Headache frequency
(ICHD- 2) CG = 5 tDCS electrode 240 minutes 30-d Pain intensity
Post (1-3)
Rapinesi et al26 Chronic migraine EG = 7 dTMS H1 coil Bilateral 100% RMT 10 12 sessions / Pre Headache frequency
(ICHD-3) CG = 7 DLPFC 4320 pulses 2-w Pain intensity
4-w VAS (0-100)
Post
Przeklasa- Migraine (IHS) EG = 30 Anodal tDCS 35 cm2 M1 2 mA NA 10 sessions / Pre Headache frequency
Muszynska CG = 20 electrode 200 minutes 30-d Pain intensity
et al31 Post NRS (0-10)

Sham tDCS was performed by adjusting ramp periods at the beginning and at the end of stimulation to mimic initial local effects of active tDCS.
CG = control group; d = day; DLPFC = dorsal lateral prefrontal cortex; dTMS = deep transcranial magnetic stimulation; EG = experimental group; FU = fellow-up; ICHD-2 =
International Classification of Headaches Disorders 2nd vision; ICHD-3 = International Classification of Headaches Disorders 3rd vision; HIS = International Headache Society;
LDLPFC = left dorsal lateral prefrontal cortex; LM1 = left primary motor cortex; M1 = primary motor cortex; m = month; NA = not available; NRS = numerical rating scale;
RMT = resting motor threshold; rTMS = repetitive transcranial magnetic stimulation; tDCS = transcranial direct current stimulation; TMS = transcranial magnetic stimulation;
V1 = visual cortex; VAS = visual analogue scale; w = week.
October 2019
Headache 1441

Table 2.—Quality Assessment of Included Studies by PEDro Scale

Study Item 1 Item 2 Item 3 Item 4 Item 5 Item 6 Item 7 Item 8 Item 9 Item 10 Total Scores

Brighina et al15 1 0 1 1 0 1 1 1 1 1 8
Teepker et al24 1 0 1 1 0 0 1 0 1 1 6
Auvichayapat et al29 1 0 1 1 0 1 1 1 1 1 8
Dasilva et al28 1 0 1 1 0 0 1 1 1 1 7
Misra et al25 1 1 1 1 1 1 1 0 1 1 9
Conforto et al27 1 1 1 1 0 1 0 0 1 1 7
Rocha et al30 1 0 1 1 0 1 0 0 1 1 6
Rapinesi et al26 1 0 1 0 0 0 1 1 1 1 6
Przeklasa-Muszynska et al31 1 0 1 0 0 0 1 1 1 1 6

Item 1: random allocation; item 2: concealment of allocation; item 3: baseline equivalence; item 4: blinding procedure (subjects); item
5: blinding procedure (therapists); item 6: blinding procedure (assessors); item 7: adequate follow-up; item 8: intention to treat analysis;
item 9: between-group statistical analysis item 10: measurement of data variability and point estimates.

Table 3.—The Results of Meta-Analysis and Sensitivity Analyses

Number of Pooled Effect Size

Analyzed Units (Hedges’ g) 95% CI P

Meta-analysis: Excitatory stimulation of M1 5 −0.935 −1.280 to −0.590 P < .001

(Headache intensity)
Subject-blinded studies 3 −0.735 −1.068 to −0.403 P < .001
Assessor-blinded studies 2 −0.764 −1.114 to −0.414 P < .001
Studies without drop out cases or with intention 4 −1.110 −1.527 to −0.693 P < .001
to treat analysis of drop out cases
Meta-analysis: Excitatory stimulation of M1 4 −0.879 −1.381 to −0.377 P = .001
(Frequency of headache attack)
Subject-blinded studies 2 −1.261 −1.631 to −0.891 P < .001
Assessor-blinded studies 2 −1.261 −1.631 to −0.891 P < .001
Studies without drop out cases or with intention 3 −0.683 −1.266 to −0.099 P = .022
to treat analysis of drop out cases
Meta-analysis: Excitatory stimulation of DLPFC 3 −1.139 −2.207 to −0.071 P = .037
(Headache intensity)
Subject-blinded studies 2 −0.795 −2.091 to 0.501 P = .229
Assessor-blinded studies 2 −0.795 −2.091 to 0.501 P = .229
Studies without drop out cases or with intention 2 −1.701 −2.569 to −0.833 P < .001
to treat analysis of drop out cases

DLPFC,15,26,27 4 studies targeted the M1,25,28,29,31 with 1
study targeting the visual cortex,30 and another 1 study
targeting the cranial vertex.24 All studies used multi-
session NIBS approach. The mean number of stimulation
sessions was 11.89, ranging from 3 sessions to 23 sessions.
Methodological Quality of Included Studies.—The
results of the methodological quality assessment
by PEDro are summarized in Table 2. The mean
score of included studies was 7.00, ranging from 6
to 9. All studies had a PEDro score  ≥  6, indicating a
high methodological quality. However, we noted that
2 studies did not apply subject blinding using a sham
stimulation control.26,31
Results of Meta-Analysis and Sensitivity Analysis.—
The results of meta-analysis and subsequent sensitivity
analysis are summarized in Table 3.
Excitatory NIBS of M1.—A total of 4 studies investi-
gated the effects of excitatory stimulation of the
1442
Fig. 2.—Meta-analysis: Excitatory stimulation of M1 (headache intensity).
Fig. 3.—Meta-analysis: Excitatory stimulation of M1 (frequency of headache attack).
M1.25,28,29,31 Meta-analysis based on these 4 studies
showed a statistically significant effect on reducing pain
intensity with a large effect size (Hedges’ g  =  −0.94;
95% confident interval [CI], −1.28 to −0.59; P < .001,
Fig. 2) with a low level of heterogeneity (I2 = 18.39%).
The overall pain reducing effect remained statistically
significant in leave-one-out sensitivity analysis (Hedges’
g from −1.11 to −0.83). The estimated effect size
remained statistically significant in our subsequent
sensitivity analyses that took into consideration the
methodological quality, with a range from −1.110 to
−0.735. Meta-analysis based on 3 studies25,29,31 showed
a significant effect on reducing the frequency of
headache attacks with a large effect size (Hedges’
g  =  −0.88; 95% CI, −1.38 to −0.38; P  =  .001, Fig. 3)
with a moderate level of heterogeneity (I2  =  57.15%),
which remained statistically significant in leave-one-out
sensitivity analysis (Hedges’ g from −1.09 to −0.68).
The estimated effect size also remained statistically
significant in our subsequent sensitivity analyses that
took into consideration the methodological quality,
with a range from −1.261 to −0.683.
October 2019
Subgroup analysis was performed to investigate
whether concurrent medicine use influences the effect
of NIBS. By removing the studies in which partici-
pants did not receive any medicine for pain,28,29 meta-
analysis showed significant effects on reducing the
headache intensity (Hedges’ g = −1.10; 95% CI, −1.69 to
−0.51; P < .001) and the frequency of headache attack
(Hedges’ g = −0.74; 95% CI, −1.43 to −0.06; P = .033),
which were similar to our primary meta-analysis.
Excitatory NIBS of DLPFC.—When investigating
the effects of excitatory stimulation of DLPFC, 3
studies were identified.15,26,27 Meta-analysis based
on these studies showed a statistically significant
effect with a large effect size on headache intensity
(Hedges’ g = −1.14; 95% CI, −2.21 to −0.07; P = .04,
Fig. 4) with a moderate level of heterogeneity
(I2  =  61.86%), but the results were not statistically
significant in leave-one-out sensitivity analysis
when excluding either Brighina et al15 or Rapinesi
et al.26 When the analysis was restricted to subject-
blinded or assessor-blinded studies, the results from
meta-analysis were not statistically significant, with a
Headache
Fig. 4.—Meta-analysis: Excitatory stimulation of DLPFC (headache intensity).
Fig. 5.—Meta-analysis: Excitatory stimulation of DLPFC (frequency of headache attack).
moderate effect size (Hedges’ g = −0.795). Moreover,
a statistically non-significant effect was noted on the
frequency of headache attacks (Hedges’ g  =  −0.64;
95% CI, −2.77 to 1.49; P  =  .56, Fig. 5), with a high
level of heterogeneity (I2  =  89.54%), although the
magnitude of effect size was considered moderate.
Inhibitory NIBS Protocols.—One study applied low-
frequency rTMS to the cranial vertex,24 which yielded
a statistically non-significant effect on headache
intensity (Hedges’ g  =  0.10; 95% CI, −0.63 to 0.83;
P = .79) as well as the frequency of headache attacks
(Hedges’ g = −0.25; 95% CI, −1.00 to 0.48; P = .50).
The magnitude of effect size was considered small.
Another study applied cathodal tDCS to the visual
cortex30 and statistically non-significant significant
results were found in headache intensity (Hedges’
g = −0.55; 95% CI, −1.58 to 0.48; P = .30) and in the
frequency of headache attacks (Hedges’ g = 0.35; 95%
CI, −0.67 to 1.36; P = .51).
DISCUSSION
Our review demonstrates that applying excitatory
NIBS on the M1 significantly reduced the frequency
1443
of headache attacks and headache intensity in
patients with migraine. This significant effect result
was robust to leave-one-out sensitivity analysis as well
as sensitivity analyses which take into consideration
methodological quality. Similarly, excitatory NIBS
of the DLPFC significantly reduced headache inten-
sity in patients with migraine. However, it did not sig-
nificantly reduce the frequency of headache attacks
in patients with migraine based on the results of our
meta-analysis. Inhibitory NIBS applied to either the
vertex or the visual cortex did not significantly change
the headache intensity and the frequency of headache
attacks in patients with migraine when compared to
sham stimulation.
The efficacy of NIBS on pain management in pa-
tients with migraine is still controversial based on prior
meta-analytic reviews. Shirahige et al32 reviewed 8 studies
and came to the conclusion that tDCS, rather than TMS,
was likely to reduce the headache intensity and the fre-
quency of migraine attacks. However, in another review
published in 2017, Lan et al33 found that single-pulse
TMS and rTMS were able to reduce the frequency of
headache attacks, based on a meta-analysis of 5 studies.
1444
We found that both of these meta-analytic studies do
not systematically investigate the efficacy of different
NIBS protocols based on the stimulation properties (ie,
excitatory or inhibitory) or the brain targets (ie, M1,
DLPFC, or visual cortex). This is contrary to the prac-
tical design of NIBS protocols for clinical applications
which use these as the main factors for consideration.
Our review separates the studies into subgroups, and
through this, we found that excitatory M1 stimulation
provides migraineurs with some benefits. The result is
in line with 2 other studies in which a favorable effect of
M1 stimulation on reducing migraine intensity and fre-
quency was found. The 2 studies were not included in our
meta-analysis because they did not report enough data
for the estimation of effect sizes.34,35 Inhibitory rTMS
via a circular coil to the vertex – which is assumed to
suppress the excitability of bilateral motor cortices –
was not superior to sham stimulation in the reduction
of headache intensity and frequency of headache at-
tack, according to one study24 included in our analysis.
Motor cortex excitability in patients with migraine is still
controversial, with some studies reporting hyperactive
interictal cortical excitability of motor cortex, whereas
others find hypoactive cortical excitability instead.36-38
However, the analgesic effect of motor cortex stimu-
lation is likely to be associated with circuit inhibition
within the somatosensory cortices,39 the recruitment of
cortico-thalamic pathway40 and the facilitation of the
top-down pain inhibitory pathway.41 Cortical excitabil-
ity, measured using motor evoked potentials alone, does
not reflect changes in functional connectivity in patients
with migraine, and these changes in connectivity are still
yet unexplored.
By performing meta-analyses on studies that stim-
ulate the DLPFC region, we found a significant effect
on reducing pain intensity but not the frequency of
headache attacks in patients with migraine. Among the
3 studies in the meta-analysis, we found that deep TMS
yielded the strongest effect on reducing pain intensity
with a large effect size (Hedges’s = −1.87). Deep TMS is
a new form of rTMS therapy that uses a H-shape coil,
which is assumed to be able to reach deeper and wider
brain regions.42 However, because this study excluded
subject- and assessor-blinding methods, the effect size
may be overestimated. In our sensitivity analysis to as-
sess the methodological quality, the overall estimation
October 2019
of effect size became statistically insignificant when
removing this study. In agreement with this, Conforto
et al27 compared the DLPFC active and sham stimu-
lation and found a superior effect favoring the sham
stimulation to the active stimulation in frequency of
headache attacks, which suggests a placebo effect
of rTMS on reducing pain intensity in migraineurs,
which contributes to significant heterogeneity in our
meta-analysis. As an appropriate blinding method is
important in controlling the placebo effect of NIBS,
it is necessary to employ a single-blinded or double-
blinded design for future studies on NIBS.
The differential effects of M1 and DLPFC stimu-
lation have not been fully confirmed. Andrade et al43
directly compared the effects of tDCS to the M1 with
that to the DLPFC, with a reported improved effect
associated with DLPFC stimulation. We cannot
include in our review due to insufficient reported data
for our meta-analysis. Interestingly, they found that
DLPFC stimulation produced a superior effect as com-
pared to stimulation of the M1, although it must be
noted that the sample size of this study was limited.
These studies suggest that the differential effects of M1
and DLPFC stimulation are of significant interest and
hence should be further investigated.
Some neurophysiological studies have found that
the visual cortex is hyperactive during the interictal pe-
riods of migraine, and hence suppressing the excitability
of the visual cortex may contribute to the reduction of
migraine pain. One of these studies, performed by
Rocha et al30 which we have included, showed an insig-
nificant trend supporting the positive effect of inhibitory
visual cortex stimulation on migraine pain. A similar
study by Antal et al,44 which was excluded from our
review because of its study design, also supported the
positive effect of inhibitory visual cortex stimulation on
migraine pain. In line with the limited number of studies
available, we were not able to draw a conclusion with
regards to the effects of inhibitory visual cortex stimu-
lation on headache in patients suffering from migraine.
LIMITATIONS
Our review is not free from limitations. First, our
meta-analysis was performed based on a limited num-
ber of articles and the estimation of effect size may
not be properly powered. Two studies with relatively
Headache
larger sample sizes 25,31 contribute to the majority of the
patients and may have undue influence on the estima-
tion of effect sizes. Hence, our conclusions should be
interpreted with caution. More studies will be needed
to confirm our present findings. Second, we were
unable to investigate potential differences in effect
between rTMS and tDCS due to the limited number
of studies that have been included. Last, although in-
cluded studies only considered migraine patients, we
cannot exclude potential overlapping headache disor-
ders among the included subjects after checking the
inclusion of each study, which might affect the valid-
ity of our results.
CONCLUSIONS
Excitatory NIBS of the M1 is likely to reduce the
headache intensity and the frequency of headache at-
tacks in patients with migraine. Our findings suggest
that therapeutic NIBS for migraine should be focused
on excitatory stimulation on the M1. Further studies
will be needed to optimize intervention parameters of
M1 stimulation in the treatment of migraine.
STATEMENT OF AUTHORSHIP
Category 1
(a) Conception and Design
Jiaqi Zhang, Yali Feng, Bingbing Zhang
(b) Acquisition of Data
Yali Feng, Bingbing Zhang
(c) Analysis and Interpretation of Data
Jiaqi Zhang, Bingbing Zhang, Yali Feng
Category 2
(a) Drafting the Manuscript
Yali Feng, Bingbing Zhang
(b) Revising It for Intellectual Content
Jiaqi Zhang, Ying Yin
Category 3
(a) Final Approval of the Completed Manuscript
Ying Yin
REFERENCES
1. Collaborators GBDH. Global, regional, and national
burden of migraine and tension-type headache,
1445
1990-2016: A systematic analysis for the Global
Burden of Disease Study 2016. Lancet Neurol.
2018;17:954-976.
2. Charles A. Migraine. N Engl J Med. 2017;377:553-561.
3. Dodick DW. Migraine. Lancet. 2018;391:1315-1330.
4. Diener HC, Charles A, Goadsby PJ, Holle D. New
therapeutic approaches for the prevention and treat-
ment of migraine. Lancet Neurol. 2015;14:1010-1022.
5. Grimsrud KW, Halker Singh RB. Emerging treat-
ments in episodic migraine. Curr Pain Headache Rep.
2018;22:61.
6. Schoenen J, Roberta B, Magis D, Coppola G.
Noninvasive neurostimulation methods for mi-
graine therapy: The available evidence. Cephalalgia.
2016;36:1170-1180.
7. Barker AT, Shields K. Transcranial magnetic stimula-
tion: Basic principles and clinical applications in mi-
graine. Headache. 2017;57:517-524.
8. Martelletti P, Jensen RH, Antal A, et al. Neuromodu-
lation of chronic headaches: Position statement from
the European Headache Federation. J Headache Pain.
2013;14:86.
9. Kobayashi M, Pascual-Leone A. Transcranial
magnetic stimulation in neurology. Lancet Neurol.
2003;2:145-156.
10. Valero-Cabre A, Amengual JL, Stengel C, Pascual-
Leone A, Coubard OA. Transcranial magnetic stim-
ulation in basic and clinical neuroscience: A compre-
hensive review of fundamental principles and novel
insights. Neurosci Biobehav Rev. 2017;83:381-404.
11. Brighina F, Cosentino G, Fierro B. Chapter 47 – Brain
stimulation in migraine. In: Lozano AM, Hallett M,
eds. Handbook of Clinical Neurology. Amsterdam:
Elsevier; 2013:585-598.
12. O’Connell NE, Marston L, Spencer S, DeSouza LH,
Wand BM. Non-invasive brain stimulation tech-
niques for chronic pain. Cochrane Database Syst Rev.
2018;4:CD008208. Available at https://www.ncbi.nlm.
nih.gov/pubmed/29652088.
13. Kurt E, Henssen D, Steegers M, et al. Motor cortex
stimulation in patients suffering from chronic neuro-
pathic pain: Summary of expert meeting and premeet-
ing questionnaire, combined with literature review.
World Neurosurg. 2017;108:254-263.
14. Lorenz J, Minoshima S, Casey KL. Keeping pain out
of mind: The role of the dorsolateral prefrontal cortex
in pain modulation. Brain. 2003;126:1079-1091.
15. Brighina F, Piazza A, Vitello G, et al. rTMS of the
prefrontal cortex in the treatment of chronic migraine:
A pilot study. J Neurol Sci. 2004;227:67-71.
1446
16. Antal A, Temme J, Nitsche MA, Varga ET, Lang N,
Paulus W. Altered motion perception in migraineurs:
Evidence for interictal cortical hyperexcitability.
Cephalalgia. 2005;25:788-794.
17. Aurora SK, Ahmad BK, Welch KM, Bhardhwaj
P, Ramadan NM. Transcranial magnetic stimula-
tion confirms hyperexcitability of occipital cortex in
migraine. Neurology. 1998;50:1111-1114.
18. Chadaide Z, Arlt S, Antal A, Nitsche MA, Lang N,
Paulus W. Transcranial direct current stimulation
reveals inhibitory deficiency in migraine. Cephalalgia.
2007;27:833-839.
19. Huang J, Cooper TG, Satana B, Kaufman DI, Cao Y.
Visual distortion provoked by a stimulus in migraine
associated with hyperneuronal activity. Headache.
2003;43:664-671.
20. Mulleners WM, Chronicle EP, Palmer JE, Koehler PJ,
Vredeveld JW. Visual cortex excitability in migraine
with and without aura. Headache. 2001;41:565-572.
21. Moher D, Liberati A, Tetzlaff J, Altman DG, Group
P. Preferred reporting items for systematic reviews
and meta-analyses: The PRISMA statement. J Clin
Epidemiol. 2009;62:1006-1012.
22. Bhogal SK, Teasell RW, Foley NC, Speechley MR. The
PEDro scale provides a more comprehensive measure of
methodological quality than the Jadad scale in stroke reha-
bilitation literature. J Clin Epidemiol. 2005;58:668-673.
23. Higgins JP. Cochrane handbook for systematic re-
views of interventions version 5.1.0 [updated March
2011]. The Cochrane Collaboration. 2011.
24. Teepker M, Hotzel J, Timmesfeld N, et al. Low-
frequency rTMS of the vertex in the prophylactic
treatment of migraine. Cephalalgia. 2010;30:137-144.
25. Misra UK, Kalita J, Bhoi SK. High-rate repetitive
transcranial magnetic stimulation in migraine pro-
phylaxis: A randomized, placebo-controlled study.
J Neurol. 2013;260:2793-2801.
26. Rapinesi C, Del Casale A, Scatena P, et al. Add-on
deep Transcranial Magnetic Stimulation (dTMS) for
the treatment of chronic migraine: A preliminary
study. Neurosci Lett. 2016;623:7-12.
27. Conforto AB, Amaro E, Jr., Goncalves AL, et al.
Randomized, proof-of-principle clinical trial of active
transcranial magnetic stimulation in chronic migraine.
Cephalalgia. 2014;34:464-472.
28. Dasilva AF, Mendonca ME, Zaghi S, et al. tDCS-
induced analgesia and electrical fields in pain-related
neural networks in chronic migraine. Headache. 2012;
52:1283-1295.
October 2019
29. Auvichayapat P, Janyacharoen T, Rotenberg A, et al.
Migraine prophylaxis by anodal transcranial direct
current stimulation, a randomized, placebo-controlled
trial. J Med Assoc Thai. 2012;95:1003-1012.
30. Rocha S, Melo L, Boudoux C, Foerster A, Araujo D,
Monte-Silva K. Transcranial direct current stimula-
tion in the prophylactic treatment of migraine based
on interictal visual cortex excitability abnormalities:
A pilot randomized controlled trial. J Neurol Sci.
2015;349:33-39.
31. Przeklasa-Muszynska A, Kocot-Kepska M, Dobro-
gowski J, Wiatr M, Mika J. Transcranial direct current
stimulation (tDCS) and its influence on analgesics
effectiveness in patients suffering from migraine head-
ache. Pharmacol Rep. 2017;69:714-721.
32. Shirahige L, Melo L, Nogueira F, Rocha S, Monte-
Silva K. Efficacy of noninvasive brain stimulation on
pain control in migraine patients: A systematic review
and meta-analysis. Headache. 2016;56:1565-1596.
33. Lan L, Zhang X, Li X, Rong X, Peng Y. The efficacy
of transcranial magnetic stimulation on migraine:
A meta-analysis of randomized controlled trails.
J Headache Pain. 2017;18:86. Available at https://thejo
urnal ofhea dache andpa in.biome dcent ral.com/track/
pdf/10.1186/s10194-017-0792-4.
34. Shehata HS, Esmail EH, Abdelalim A, et al.
Repetitive transcranial magnetic stimulation ver-
sus botulinum toxin injection in chronic migraine
prophylaxis: A pilot randomized trial. J Pain Res.
2016;9:771-777.
35. Misra UK, Kalita J, Tripathi G, Bhoi SK. Role of beta
endorphin in pain relief following high rate repetitive
transcranial magnetic stimulation in migraine. Brain
Stimul. 2017;10:618-623.
36. Neverdahl JP, Omland PM, Uglem M, Engstrom
M, Sand T. Reduced motor cortical inhibition in
migraine: A blinded transcranial magnetic stimulation
study. Clin Neurophysiol. 2017;128:2411-2418.
37. Khedr EM, Ahmed MA, Mohamed KA. Motor and
visual cortical excitability in migraineurs patients with
or without aura: Transcranial magnetic stimulation.
Neurophysiol Clin. 2006;36:13-18.
38. Bohotin V, Fumal A, Vandenheede M, Bohotin C,
Schoenen J. Excitability of visual V1-V2 and motor
cortices to single transcranial magnetic stimuli in
migraine: A reappraisal using a figure-of-eight coil.
Cephalalgia. 2003;23:264-270.
39. Garcia-Larrea L, Peyron R, Mertens P, et al.
Positron emission tomography during motor
Headache
cortex stimulation for pain control. Stereotact Funct
Neurosurg. 1997;68:141-148.
40. Briggs F, Usrey WM. Emerging views of corti-
cothalamic function. Curr Opin Neurobiol. 2008;18:
403-407.
41. Peyron R, Faillenot I, Mertens P, Laurent B, Garcia-
Larrea L. Motor cortex stimulation in neuropathic
pain. Correlations between analgesic effect and
hemodynamic changes in the brain. A PET study.
Neuroimage. 2007;34:310-321.
42. Roth Y, Amir A, Levkovitz Y, Zangen A. Three-
dimensional distribution of the electric field induced
in the brain by transcranial magnetic stimulation
using figure-8 and deep H-coils. J Clin Neurophysiol.
2007;24:31-38.
1447
43. Andrade SM, de Brito Aranha REL, de Oliveira EA,
et al. Transcranial direct current stimulation over
the primary motor vs prefrontal cortex in refractory
chronic migraine: A pilot randomized controlled trial.
J Neurol Sci. 2017;378:225-232.
44. Antal A, Kriener N, Lang N, Boros K, Paulus W.
Cathodal transcranial direct current stimulation of
the visual cortex in the prophylactic treatment of
migraine. Cephalalgia. 2011;31:820-828.
SUPPORTING INFORMATION
Additional supporting information may be found
in the online version of this article at the publisher’s
web site.