Research Submission

Reduced Threshold for Inhibitory Homeostatic Responses inMigraine Motor Cortex? A tDCS/TMS Study

Giuseppe Cosentino, MD; Filippo Brighina, MD; Simona Talamanca, MD; Piera Paladino, MD;Simone Vigneri, MD; Roberta Baschi, MD; Serena Indovino, MD; Simona Maccora, MD;

Enrico Alfonsi, MD; Brigida Fierro, MD

Background and Objective.—Neurophysiological studies in migraine have reported conflicting findings of either corticalhyper- or hypoexcitability. In migraine with aura (MwA) patients, we recently documented an inhibitory response tosuprathreshold, high-frequency repetitive transcranial magnetic stimulation (hf-rTMS) trains applied to the primary motorcortex, which is in contrast with the facilitatory response observed in the healthy subjects. The aim of the present study was tosupport the hypothesis that in migraine, because of a condition of basal increased cortical responsivity, inhibitory homeostatic-like mechanisms of cortical excitability could be induced by high magnitude stimulation. For this purpose, the hf-rTMS trainswere preconditioned by transcranial direct current stimulation (tDCS), a noninvasive brain stimulation technique able tomodulate the cortical excitability state.

Methods.—Twenty-two MwA patients and 20 patients with migraine without aura (MwoA) underwent trains of 5-Hzrepetitive transcranial magnetic stimulation at an intensity of 130% of the resting motor threshold, both at baseline and afterconditioning by 15 minutes of cathodal or anodal tDCS. Motor cortical responses to the hf-rTMS trains were compared withthose of 14 healthy subjects.

Results.—We observed abnormal inhibitory responses to the hf-rTMS trains given at baseline in both MwA and MwoApatients as compared with the healthy subjects (P < .00001). The main result of the study was that cathodal tDCS, which reducesthe cortical excitability level, but not anodal tDCS, which increases it, restored the normal facilitatory response to the hf-rTMStrains in both MwA and MwoA.

Conclusions.—The present findings strengthen the notion that, in migraine with and without aura, the threshold forinducing inhibitory mechanisms of cortical excitability might be lower in the interictal period. This could represent a protectivemechanism counteracting cortical hyperresponsivity. Our results could be helpful to explain some conflicting neurophysiologi-cal findings in migraine and to get insight into the mechanisms underlying recurrence of the migraine attacks.

Key words: migraine, repetitive transcranial magnetic stimulation, transcranial direct current stimulation, homeostatic plastic-ity, metaplasticity

(Headache 2014;54:663-674)

Cortical homeostatic plasticity refers to physi-ological mechanisms acting in the human brain tostabilize cortical activity within a useful dynamicrange.1 In recent years, homeostatic plasticity hasbeen non-invasively investigated in the human cortexby combining different brain stimulation techniquessuch as repetitive transcranial magnetic stimulation(rTMS) and transcranial direct current stimulation

From the Department of Experimental Biomedicine and Clini-cal Neurosciences (BioNeC), University of Palermo, Palermo,Italy (G. Cosentino, F. Brighina, S. Talamanca, P. Paladino, S.Vigneri, R. Baschi, S. Indovino, S. Maccora, and B. Fierro);Spinal and Cranial Reflexes Laboratory, National Institute ofNeurology Foundation “C. Mondino,” Pavia, Italy (E. Alfonsi).

Address all correspondence to F. Brighina, Via Gaetano LaLoggia, 1, 90129 Palermo, Italy, email: filippo.brighina@unipa.it

Accepted for publication July 30, 2013. Conflict of Interest: None.

ISSN 0017-8748doi: 10.1111/head.12249

Published by Wiley Periodicals, Inc.Headache© 2013 American Headache Society

663

(tDCS).2-4 In last decades, different rTMS paradigmshave been widely used to test synaptic plasticity in thehuman cortex.5 tDCS, which has been more recentlyintroduced, consists of the application of weak directcurrents to the scalp which is capable of eliciting cor-tical excitability changes.6 The nature of this modula-tion depends on tDCS polarity. Anodal tDCSincreases excitability, whereas cathodal tDCS dimin-ishes it.7

One of the main principles of homeostatic plas-ticity says that the increase in cortical activationlowers the threshold for inducing long-term depres-sion (LTD) responses, while the decrease in corticalactivity raises it.8,9 Accordingly, Lang et al2 showedthat the well-known long-term potentiation-likeafter-effects induced in the motor cortex by a sessionof high-frequency rTMS (hf-rTMS) could turn intoLTD-like responses when hf-rTMS was preceded by asession of anodal tDCS that increases the corticalexcitability state.

Evidence has been also provided that brief trainsof hf-rTMS, when combined with “facilitatory”anodal tDCS, may probe inhibitory homeostaticmechanisms of cortical excitability mainly acting atthe presynaptic level of glutamatergic intracorticalcircuits.10 Indeed, in healthy subjects, the progressivepotentiation of the motor evoked potentials (MEPs)normally elicited by the trains11,12 could turn into inhi-bition after experimentally increasing the corticalexcitability state.10

Recently, we showed in migraine with aura(MwA) patients that suprathreshold (130% of theresting motor threshold [RMT]) hf-rTMS trains of10 stimuli elicit a paradoxical inhibitory MEPsresponse throughout the train, which resembles thatobserved in the healthy subjects after anodal tDCSpreconditioning.13 We supposed that in a conditionof cortical hyperresponsivity, as evidenced in mig-raine by different neurophysiological studies,14-16 thethreshold for inhibitory homeostatic responses couldslide down as a compensatory mechanism to preventexcessive and runaway excitation in response to highmagnitude stimuli. The aim of this study was to lendsupport to this hypothesis by combining in MwApatients the suprathreshold hf-rTMS trains withtDCS, namely with the same paradigm as we previ-

ously used in the healthy subjects. We assumed that apreconditioning session of cathodal tDCS couldrestore the facilitatory response to the hf-rTMStrains in the migraine motor cortex. Indeed, cathodaltDCS transiently decreases cortical excitability andthus, according to the rules of cortical homeostaticplasticity, could raise the threshold for inducinginhibitory homeostatic responses.2,17 Results werealso compared with respect to migraine without aura(MwoA) patients.

SUBJECTS AND METHODSSubjects.—Twenty-two patients affected by MwA

(7 males/15 females, mean age 36 years ± 10 standarddeviation [SD]), 20 patients affected by MwoA (5males/15 females, mean age 36.4 years ± 9.8 SD), and14 healthy subjects (5 males/9 females, mean age31.1 years ± 6.6 SD) participated in the study.Patients were recruited from the Headache Outpa-tient Service of the Neurology Department at theUniversity of Palermo, Italy. Diagnosis of MwA andMwoA was based on the diagnostic criteria of theHeadache Classification Subcommittee of the Inter-national Headache Society (2004). All the patientssuffering from MwA experienced visual aura in atleast 50% of their attacks. None of the MwA patientshad sensory or hemiplegic aura symptoms. Patientswere examined in the interictal period (at least 48hours before and after attacks:absence of attacks afterrecording being verified by means of a telephone call)and, at the time of the experiments, they were nottaking any prophylactic drug for a period of at least 3months. To avoid nonspecific effects on cortical excit-ability, all the female subjects were not examinedduring the menstrual phase. Before enrollment, all thesubjects were checked for contraindications totranscranial magnetic stimulation (TMS) and tDCS18

and gave their written informed consent to participate.The local ethics committee approved the experimentalprocedures. Patients’ demographic and clinical dataare presented in Table 1.

Stimulation Procedures.—All the subjects werecomfortably seated on a chair and told to be asrelaxed as possible. They wore a tight-fitting plasticswimming cap to mark the optimum stimulation siteand ensure optimum coil placement. Electromyogra-

664 April 2014

phy (EMG) signals were recorded from the rightabductor pollicis brevis (APB) muscle using 0.9-cmdiameter Ag-AgCl surface electrodes placed 3 cmapart over the belly and tendon of the muscle. TheEMG activity was recorded with a bandpass of10-1000 Hz and a display gain ranging from 50 to1000 μV/cm. EMG signals were collected, averaged,and analyzed offline. Focal TMS was delivered overthe hand motor cortex of the left hemisphere using afigure-of-8 coil connected to a monophasic Cadwellhigh speed magnetic stimulator (Cadwell Laborato-ries, Kennewick,WA, USA).The stimulating coil withposteroanterior orientation was placed over theoptimal site for eliciting responses in the contralateraltarget muscle.19 The RMT for eliciting responses in therelaxed APB muscle was defined as the minimumintensity of stimulation needed to produce responsesof 50 μV in at least 50% of 10 trials.The subjects weregiven audiovisual feedback of EMG activity to helpmaintain complete muscle relaxation.The coil positionwas continuously monitored throughout the experi-ment in order to keep it constant.The examiners wereblind to the patients’ diagnosis (MwoA vs MwA) atthe time of the evaluation. Stimulation was performedfollowing safety guidelines.20,21 Continuous tDCS wasdelivered through a pair of electrodes in a 5 × 7 cmwater-soaked synthetic sponge using a battery-drivenconstant current stimulator (Eldith DC-Stimulator,neuroConn, Ilmenau, Germany). The first electrodewas positioned over the motor hotspot of the rightABP muscle, as revealed by TMS. The second elec-trode was placed above the contralateral orbit. tDCSpolarity refers to the electrode placed over the leftprimary motor cortex. Currents were delivered for 15minutes at an intensity of 1.5 mA and were ramped upor down over the first and last 8 seconds of stimulation.

Experimental Paradigm and Measurements.—Allthe patients and healthy subjects underwent a base-line session in which 6 rTMS trains of 10 stimuli weredelivered at 5 Hz and 130% RMT intensity to the leftprimary motor hand area.

Fourteen of the 22 MwA patients and 12 of the20 MwoA patients enrolled were randomly selectedto participate in the main experiment, which wasdesigned to explore the preconditioning effect ofcathodal tDCS on the corticospinal response to thehf-rTMS trains. This consisted of 2 other sessions inwhich the 6 rTMS trains were delivered, in a counter-balanced order, immediately or 20 minutes after theend of cathodal tDCS preconditioning. In these pre-conditioned sessions, the rTMS trains were given, ineach patient, at an intensity of the stimulator outputadjusted to evoke MEP amplitudes in the range of theaveraged first MEP size in the rTMS trains performedin the baseline condition.This was done because it hasbeen shown that tDCS preconditioning may inducechanges in MEPs size elicited by suprathreshold TMSstimuli,6 in order to prevent any differences in the firstMEP amplitudes from interfering with the subse-quent motor cortical responses to the trains.

Eight of the 22 MwA patients and 8 of the 20MwoA patients were randomly selected to partici-pate in a supplementary experiment. This wasdesigned to explore the preconditioning effect ofanodal tDCS on the motor cortical response to therTMS trains. It consisted of a session in which anodaltDCS was applied to the left primary motor corteximmediately before the six 5-Hz rTMS trains. As inthe main experiment, the intensity of the 5-Hz rTMStrains was adjusted after anodal tDCS precondition-ing so that the amplitude of the first MEP evoked bythe trains matched the MEP amplitude at baseline.

Table 1.—Demographic and Clinical Characteristics of the Enrolled Patients: Mean (SD) or Count

Age (years)Headache

history (years)Attack frequency

(n/month)Resting motor

threshold

Migraine with aura(MwA)

Main exp. (n = 14/22) 36.0 (10) 14.5 (9) 4.1 (2.3) 61.8 (8.6)Suppl. exp. (n = 8/22) 35.6 (13.3) 15.4 (11.3) 3.9 (2.4) 63.6 (8.4)

Migraine without aura(MwoA)

Main exp. (n = 12/20) 35.6 (9.9) 16.2 (8.5) 4.1 (1.1) 65.3 (11.6)Suppl. exp. (n = 8/20) 35.7 (9.7) 14.6 (7.1) 3.7 (1.5) 57.9 (7.4)

Headache 665

In the female patients, the different evaluationswere performed always in the same period of themenstrual cycle.22 In the healthy subjects, experimen-tal assessments were made only at baseline becausethe aim was to evaluate to which extent the responseof the migraine motor cortex to the rTMS trains coulddiffer from the normal one at baseline.

In all sessions, the rTMS trains were applied witha 2-minute intertrain interval on subjects at rest. Toevaluate changes in MEP size during the rTMS trains,for each subject, MEP amplitudes were calculatedpeak-to-peak from single traces of the 6 trains andthen averaged according to their position in the train.Intervals of at least 1 week were always allowed toelapse between different experimental sessions in asingle subject.

Statistical Analysis.—A 2-way analysis of variance(ANOVA) with “group” (3 levels: MwA, MwoA, andhealthy subjects) as between-subjects factor and“number of stimuli” (10 levels) as within-subjectsfactor was performed to compare the responses withthe rTMS trains given at baseline between patientsand healthy subjects.

In the main experiment, changes in MEP ampli-tude (from the 1st to the 10th MEP) during the 5-HzrTMS trains were compared using a 3-way repeated-measures ANOVA with “group” (2 levels: MwA andMwoA patients) as between-subjects factor, “condi-tion” (3 levels: baseline, immediately after, and 20minutes after cathodal tDCS), and “number ofstimuli” in the train (10 levels) as within-subjectsfactors.

Three-way ANOVA was also performed for thesupplementary experiment, again with “group” (2levels: MwA and MwoA patients) as between-subjects factor, “condition” (2 levels: baseline andpost-anodal tDCS), and “number of stimuli” (10levels) as the within-subjects factors.

In each condition, ANOVA included the MEPvalues obtained by averaging MEP amplitudes overthe 6 trains according to their position in the train (ie,the 1st, 2nd. . .10th averaged MEP values). In addi-tion, for each experimental session, we pooled the 10MEP amplitudes of each train together to analyzechanges in the mean MEP amplitude of each trainfrom train 1 to train 6 by using 1-way ANOVA.

Duncan’s test was used for post hoc analysis.Paired sample t-test was used to compare the stimulusintensity used at baseline with respect to that adjustedafter tDCS conditioning in each experimentalsession. For all analyses, the level of statistical signifi-cance was set at P < .05.

RESULTSAll the subjects completed the planned cortical

excitability measurements. The experimental proce-dures were well tolerated, and no adverse effects werereported. Patients who reported migraine attackswithin the 48 hours after the experimental procedureswere reassessed in a subsequent session in order toensure that the electrophysiological measurementswere always performed during an interictal phase.t-test showed no significant differences in averageage, RMT, and clinical parameters (headache history,attack frequency, severity, and duration of theattacks) between MwA and MwoA patients andbetween patients enrolled in the different experimen-tal paradigms (Table 1). No significant differences indemographic parameters and RMT were foundbetween patients and control subjects.

When comparing the intensity of the stimulatoroutput used to apply the rTMS trains between base-line and after cathodal tDCS preconditioning, weobserved a significant increase in the stimulationintensity we used immediately after the precondition-ing: from 80.5 ± 11.2 to 84.1 ± 9.3% of the maximumstimulator output in MwA patients (P < .002) andfrom 84.9 ± 15.1 to 85.6 ± 14.4% in MwoA patients(P < .05). Conversely, in the supplementary experi-ment, the intensity we used after anodal tDCSwas significantly decreased: from 80.9 ± 11.1 to77.3 ± 9.9% in MwA patients (P < .01) and from75.2 ± 9.7 to 72.7 ± 8.1% in MwoA patients (P < .05).The mean MEP amplitude of each train from train 1to train 6 was not statistically different for all theexperimental sessions.

ANOVA used to compare the response with therTMS trains given at baseline between the 2 groups ofpatients, and healthy subjects showed a significanteffect of “group” (F(2, 53) = 5,01, P < .02) and “numberof stimuli” (F(9, 477) = 3,25, P < .001) and a significantinteraction between “group” and “number of stimuli”

666 April 2014

(F(18, 477) = 5,74, P < .00001) (Fig. 1). Post hoc analysisshowed a facilitatory response with increased MEPsize from the 3rd response as compared with the 1stone in the healthy subjects (P < .05). Instead, the sizeof the 5th MEP response for MwA patients and thatof all the MEP responses following the first forMwoA patients was significantly reduced comparedwith the 1st one (P < .05).

When comparing, vs the baseline condition,changes in MEP size during the 5-Hz rTMS trainsdelivered with a 0-minute or 20-minute delay fromcathodal tDCS, ANOVA showed a significant effectof “condition” (F(2, 48) = 8,37, P < .001) and “number ofstimuli” (F(9, 216) = 3,71, P < .001) and a significantinteraction between “condition” and “number ofstimuli” (F(18, 432) = 4,04, P < .00001); no significantinteraction between “group,” “condition,” and“number of stimuli” was found (Figs. 2 and 3). Posthoc analysis showed that in the baseline condition,the size of the 5th and 6th MEPs in MwA and the sizeof all the MEPs following the first in MwoA signifi-cantly decreased as compared with the 1st response(P < .05). Conversely, a progressive MEP facilitationwas seen during the trains delivered both immedi-ately and 20 minutes after tDCS in both MwA andMwoA patients. Post hoc analysis showed in MwApatients a significant increase of the 9th and 10th

MEP size (P < .05) immediately after cathodal tDCSwith respect to the first MEP, while a significant facili-tation from the 8th to the 10th response (P < .05) wasobserved at 20-minute delay. In MwoA patients, weobserved a significant increase of the 10th response(P < .02) immediately after tDCS and a facilitationfrom the 8th to the 10th response (P < .05) at20-minute delay. In the supplementary experiment,the baseline motor cortical inhibition during thetrains did not change after anodal tDCS in both MwAand MwoA. Indeed, ANOVA showed only a signifi-cant effect of “number of stimuli” (F(9, 126) = 7,71,P < .00001) (Fig. 4).

DISCUSSIONIn the present work, first, we confirm the previ-

ously observed inhibitory response of the primarymotor cortex to brief hf-rTMS trains given at 130%RMT in MwA patients as compared with healthysubjects.13 An inhibitory response to the same rTMSparadigm was also observed in MwoA patients. Then,we evaluate the modulatory effect of tDCS on motorcortical excitability as assessed by the hf-rTMS para-digm. In recent years, tDCS has evolved as an excel-lent tool to noninvasively modulate the corticalexcitability state in vivo.6 The effects of tDCS on cor-tical activity are located intracortically because the

Fig 1.—Motor evoked potentials (MEPs) elicited by repetitive transcranial magnetic stimulation (rTMS) trains delivered at anintensity of 130% resting motor threshold (RMT) in baseline condition in migraine with aura (MwA) patients, migraine withoutaura (MwoA) patients, and healthy subjects. A progressive MEPs potentiation throughout the rTMS train may be observed in thehealthy subjects. Conversely, both MwA and MwoA patients show a similar inhibitory MEPs response. The mean MEPs size areexpressed as percentage of the first MEP size in train. Error bars indicate standard error of means (SE).

Headache 667

excitability of the cortical-spinal tract remainsunchanged.7 The effects during short-lasting tDCSare likely generated by polarity-specific shifts of theresting membrane potential in both inhibitory andfacilitatory intracortical neurons, while the formationof after-effects critically depends on modulation ofglutamatergic neurotransmission and N-methyl-D-aspartate (NMDA) receptors efficacy.6,23 Main findingof the present study was that cathodal tDCS, whichdecreases cortical excitability, may restore the normalfacilitatory response to the rTMS trains in both MwAand MwoA patients. Conversely, “facilitatory” anodaltDCS preconditioning was not found to interfere with

the baseline inhibitory response. These resultssupport the hypothesis that in migraine, the thresholdfor inhibitory homeostatic mechanisms regulatingcortical excitability could be lower to prevent exces-sive cortical activation in response to high magnitudeof stimulation, as a possible compensatory mechanismwith regard to the cortical hyperresponsivity to variousendogenous and exogenous stimuli.13,24 As an alterna-tive explanation, we could suppose that in migraine,in the interictal period, a population of intracorticalinhibitory interneurons could be hyperresponsive tohigh intensity of hf-rTMS so leading to paradoxicalinhibitory responses. Accordingly, the restoration of a

Fig 2.—Main experiment. Motor evoked potentials (MEPs) elicited by repetitive transcranial magnetic stimulation (rTMS) trainsdelivered at an intensity of 130% resting motor threshold (RMT) in baseline condition and after delays of 0 and 20 minutes fromthe end of cathodal transcranial direct current stimulation (tDCS) preconditioning. Migraine with aura (MwA) and migrainewithout aura (MwoA) patients are depicted above and below, respectively. In both MwA and MwoA patients, the inhibitory MEPsresponse observed in the baseline condition was replaced by a normal MEPs potentiation after cathodal tDCS. Normalization ofMEPs response is more evident at 20-minute delay. The mean MEPs size are expressed as percentage of the first MEP size in train.Error bars indicate standard error of means (SE).

668 April 2014

normal facilitatory response after cathodal tDCScould be attributed to hyperpolarization of hyperex-citable inhibitory interneurons induced by the stimu-lation itself. This idea, however, is not supported byfindings of reduced intracortical inhibition inmigraine25-27 and by evidence that tDCS-inducedafter-effects are mainly glutamate-dependent.6,23 Inaddition, we previously showed that the cortical silentperiod, a measure of GABAergic intracortical activ-ity, normally increases during the hf-rTMS trains inmigraineurs.13

Several neurophysiological findings support theconcept that abnormal glutamatergic neurotransmis-sion could be at the basis of cortical hyperresponsivityin migraine, eg, increased intracortical facilitation topaired-pulse TMS,14 increased slope of the input–output curves,16 and greater increasing in MEPs facili-

tation by brief trains of hf-rTMS given at 110%13 and120% RMT.15 Moreover, evidence of glutamatergichyperresponsivity in migraine comes from biologi-cal28,29 and neuroimaging30 studies. It remains unclear,however, whether glutamatergic hyperresponsivityrepresents a primitive pathogenetic mechanism inmigraine or it is a consequence of an impairment inthe activity of inhibitory intracortical circuits25-27 orof a reduced cortical preactivation level.24,31 Our find-ings are in line with the idea that glutamatergicintracortical circuits could play a pathogenetic role“per se” based on 2 considerations. First, deficit ofintracortical inhibition is not suitable to explainthe inhibitory MEPs response observed during thehf-rTMS trains at 130%. Second, the finding thatanodal tDCS preconditioning did not significantlyaffect the response to the rTMS trains weakens thesuggestion that the baseline inhibitory response isdue to reduced preactivation of the motor cortex bysubcortical structures. Indeed, were this the case, wemight have expected to see a restoration of thenormal facilitatory response to the rTMS trains byanodal rather than cathodal tDCS preconditioning. Inour opinion, however, findings of reduced inhibitoryintracortical activity, and experimental evidence thatimpaired long-term synaptic plasticity could berelated to a lower cortical preactivation level inmigraine,32 do not conflict with the present results.Indeed, on the one hand, the impairment in inhibitoryintracortical activity could be interpreted as due to afailure of GABAergic intracortical neurons in inhib-iting hyperresponsive glutamatergic circuits. Thispoint of view is supported by evidence that TMS pro-tocols testing intracortical inhibition carry the risk tomeasure cortical inhibition contaminated by activityof facilitatory intracortical neurons.33 On the otherhand, downregulation of the thalamus and brainstemnuclei could represent a compensatory mechanismtoward cortical hyperresponsivity mediated bycortico-subcortical feedback loops.34,35

Although the molecular mechanisms involved inthe observed responses cannot be directly derivedfrom our results, presynaptic mechanisms regulatingglutamate release, which are indirectly evaluated bythe hf-rTMS trains,36-39 should be taken into account.Indeed, the presynaptic terminal of the glutamatergic

Fig 3.—Representative examples of motor evoked potentials(MEPs) produced in abductor pollicis brevis (APB) muscle byrepetitive transcranial magnetic stimulation (rTMS) trains at130% of resting motor threshold (RMT) in baseline (A) andimmediately (B) and 20 minutes (C) after the end of cathodaltranscranial direct current stimulation (tDCS) conditioning in amigraine with aura (MwA) patient at rest. Notice the inhibitoryresponse observed in the baseline condition (A), with reducedamplitude of the MEPs following the first one in the train, andthe facilitatory responses recorded in the same patient aftercathodal tDCS preconditioning (B, C). The numbers on the left(1, 5, 10) indicate the number of pulses during the rTMS trains.

Headache 669

neurons represents a crucial site for homeostaticregulation of cortical excitability because presynapticmechanisms are involved in both short-termpotentiation-like responses40,41 and cortical homeo-static plasticity.42,43 In particular, it has been shownthat mechanisms regulating Ca2+ currents throughpresynaptic high-voltage-activated (HVA) Ca2+channels may be responsible both for facilitation andinhibition of glutamate release, respectively in rela-tion to lower and higher levels of activity at the pre-synaptic terminal.44-46

Interestingly, we observed a greater facilitatoryresponse to the rTMS trains delivered after a20-minute delay than to those delivered immediatelyafter the tDCS preconditioning. This could be due toreduction in the conductance of sodium and calciumchannels induced by the cathodal tDCS, which maybe responsible for decreased facilitation of glutamaterelease.47 This effect occurs during the trains deliveredimmediately after the preconditioning stimulationwhich increased the threshold for inhibitory homeo-static mechanisms. Instead, at 20 minutes post-tDCSpreconditioning, it is conceivable that reduction oftDCS-induced hyperpolarization allows the mecha-nisms of glutamatergic facilitation to act in a morepronounced fashion. This fits well with findings in

healthy subjects showing a reduced slope of MEPpotentiation during hf-rTMS trains delivered imme-diately after cathodal tDCS.10

The concept of abnormal pattern of homeostaticplasticity in migraine is in line with findings by Antalet al48 and Sándor et al49 in the motor and visual cor-tices, respectively, and it could be helpful to explainsome conflicting neurophysiological findings. Forinstance, the paradoxical inhibitory effect of a sessionof hf-rTMS on migraine motor cortical excitability50

and its ability to restore habituation to visual evokedpotentials in migraine51 have been differently inter-preted. The first has been explained as consequenceof normalization of impaired intracortical inhibitionby hf-rTMS, while the second as due to increasing inthe cortical activation level by the trains in a basalcondition of abnormal reduced cortical preactivation.In our opinion, both findings could be alternativelyreinterpreted as consequence of homeostatic inhibi-tory mechanisms of cortical hyperresponsivityinduced by the stimulation itself.

It has been supposed that an abnormal regulationof cortical homeostatic plasticity could be involved inthe susceptibility to the migraine attack.48 Our find-ings seem to suggest the concept that, in the interictalperiod, inhibitory homeostatic mechanisms of cortical

Fig 4.—Supplementary experiment. Motor evoked potentials (MEPs) elicited by repetitive transcranial magnetic stimulation(rTMS) trains delivered at an intensity of 130% resting motor threshold (RMT) in baseline condition and immediately after anodaltranscranial direct current stimulation (tDCS) preconditioning in migraine with aura (MwA) and migraine without aura (MwoA)patients. A similar inhibitory MEPs response throughout the rTMS train is observed in the baseline condition and after anodaltDCS preconditioning. The mean MEPs size are expressed as percentage of the first MEP size in train. Error bars indicate standarderror of means (SE).

670 April 2014

excitability could be more active to counteract corti-cal hyperresponsivity, rather than due to a primitivealteration. Notwithstanding, it is tempting to specu-late that a transient increase in the threshold forinhibitory homeostatic mechanisms, possibly relatedto physiological fluctuations in the cortical activa-tion level, could allow an unchallenged corticalhyperresponsivity to lead to a critical excitabilitylevel triggering the attack. However, repeated assess-ments of cortical excitability and homeostaticresponses throughout the migraine cycle will be nec-essary to address this intriguing issue.

Another important finding worthy to be dis-cussed refers to the similarity of the cortical responseto the rTMS trains between MwA and MwoApatients, both in baseline and after tDCS precondi-tioning.This suggests that, at least in the motor cortex,migraine cortical “disexcitability” could affecthomeostatic plasticity in a similar manner in the 2groups of patients. This datum is apparently in con-trast with finding by Conte et al15 showing increasedmotor cortical response to 120% high-frequencyrTMS trains in MwA but not in MwoA patients. Sucha discrepancy may be linked both to the lower stimu-lation intensity used by the authors that could be notenough to activate inhibitory homeostatic mecha-nisms, and to different clinical features of the enrolledsubjects, eg, the lower attack frequency of the MwApatients.

Our study has limitations that are worth noting.The number of the subjects is relatively low, and nopatients with nonvisual aura were enrolled into thestudy. Thus, we cannot exclude that patients withaura symptoms because of involvement of moreanterior cortical areas (ie, the precentral and post-central gyri) could have different responses to therTMS trains. Then, we evaluated patients only in theinterictal period. This does not allow us to draw anydefinite conclusions on the hypothesized role ofcortical homeostatic plasticity in the recurrence ofthe migraine attacks. Finally, although it was neces-sary to adjust the rTMS stimulation intensity aftertDCS preconditioning to ensure similar first MEPamplitudes in the rTMS trains, this artifice repre-sents a possible bias of the study that is worth ofmentioning.

In conclusion, in the present study, we get insightinto the mechanisms responsible for homeostaticregulation of cortical excitability in migraine, suggest-ing that the threshold for inhibitory homeostaticresponses could be lower in the interictal period. Thispoint of view could be helpful to understand thepathophysiological mechanisms at the basis of theepisodic nature of migraine and to interpret con-flicting neurophysiological findings. Moreover, ourresults could provide indirect evidence of a possibleglutamatergic dysfunction because of abnormalmodulation of the presynaptic HVA Ca2+ channels inmigraine. This, however, remains speculative, andmolecular studies investigating presynaptic regula-tory mechanisms of glutamatergic neurotransmissionin migraine are worth to be explored.

STATEMENT OF AUTHORSHIP

Category 1(a) Conception and Design

Giuseppe Cosentino, Filippo Brighina, BrigidaFierro

(b) Acquisition of DataGiuseppe Cosentino, Simona Talamanca, PieraPaladino, Simone Vigneri, Roberta Baschi,Serena Indovino, Simona Maccora

(c) Analysis and Interpretation of DataGiuseppe Cosentino, Filippo Brighina, SimonaTalamanca, Piera Paladino, Simone Vigneri,Enrico Alfonsi, Brigida Fierro

Category 2(a) Drafting the Manuscript

Giuseppe Cosentino(b) Revising It for Intellectual Content

Filippo Brighina, Simona Talamanca, PieraPaladino, Simone Vigneri, Roberta Baschi,Serena Indovino, Simona Maccora, EnricoAlfonsi, Brigida Fierro

Category 3(a) Final Approval of the Completed Manuscript

Giuseppe Cosentino, Filippo Brighina, SimonaTalamanca, Piera Paladino, Simone Vigneri,Roberta Baschi, Serena Indovino, SimonaMaccora, Enrico Alfonsi, Brigida Fierro

Headache 671

REFERENCES1. Bienenstock EL, Cooper LN, Munro PW. Theory

for the development of neuron selectivity: Orienta-tion specificity and binocular interaction in visualcortex. J Neurosci. 1982;2:32-48.

2. Lang N, Siebner HR, Ernst D, et al. Preconditioningwith transcranial direct current stimulation sensi-tizes the motor cortex to rapid-rate transcranialmagnetic stimulation and controls the direction ofafter-effects. Biol Psychiatry. 2004;56:634-639.

3. Siebner HR, Lang N, Rizzo V, et al. Preconditioningof low-frequency repetitive transcranial magneticstimulation with transcranial direct current stimula-tion: Evidence for homeostatic plasticity in thehuman motor cortex. J Neurosci. 2004;24:3379-3385.

4. Schabrun SM, Chipchase LS, Zipf N, ThickbroomGW, Hodges PW. Interaction between simultane-ously applied neuromodulatory interventions inhumans. Brain Stimul. 2013;6:624-630.

5. Hoogendam JM, Ramakers GM, Di Lazzaro V.Physiology of repetitive transcranial magneticstimulation of the human brain. Brain Stimul. 2010;3:95-118.

6. Nitsche MA, Seeber A, Frommann K, et al. Modu-lating parameters of excitability during and aftertranscranial direct current stimulation of the humanmotor cortex. J Physiol. 2005;568(Pt 1):291-303.

7. Nitsche MA, Paulus W. Excitability changes inducedin the human motor cortex by weak transcranialdirect current stimulation. J Physiol. 2000;527(Pt3):633-639.

8. Huang YY, Colino A, Selig DK, Malenka RC. Theinfluence of prior synaptic activity on the inductionof long-term potentiation. Science. 1992;255:730-733.

9. Kirkwood A, Rioult MC, Bear MF. Experience-dependent modification of synaptic plasticity invisual cortex. Nature. 1996;381:526-528.

10. Cosentino G, Fierro B, Paladino P, et al.Transcranial direct current stimulation precondi-tioning modulates the effect of high-frequencyrepetitive transcranial magnetic stimulation in thehuman motor cortex. Eur J Neurosci. 2012;35:119-124.

11. Pascual-Leone A, Valls-Solé J, Wassermann EM,Hallett M. Responses to rapid-rate transcranialmagnetic stimulation of the human motor cortex.Brain. 1994;117(Pt 4):847-858.

12. Berardelli A, Inghilleri M, Rothwell JC, et al. Facili-tation of muscle evoked responses after repetitivecortical stimulation in man. Exp Brain Res. 1998;122:79-84.

13. Brighina F, Cosentino G, Vigneri S, et al. Abnormalfacilitatory mechanisms in motor cortex of migrainewith aura. Eur J Pain. 2011;15:928-935.

14. Siniatchkin M, Kröner-Herwig B, Kocabiyik E,Rothenberger A. Intracortical inhibition and facili-tation in migraine – a transcranial magnetic stimu-lation study. Headache. 2007;47:364-370.

15. Conte A, Barbanti P, Frasca V, et al. Differences inshort-term primary motor cortex synaptic potentia-tion as assessed by repetitive transcranial magneticstimulation in migraine patients with and withoutaura. Pain. 2010;148:43-48.

16. Cosentino G, Fierro B, Vigneri S, et al. Impairedglutamatergic neurotransmission in migraine withaura? Evidence by an input-output curves trans-cranial magnetic stimulation study. Headache. 2011;51:726-733.

17. Jedlicka P. Synaptic plasticity, metaplasticityand BCM theory. Bratisl Lek Listy. 2002;103:137-143.

18. Keel JC, Smith MJ, Wassermann EM. A safetyscreening questionnaire for transcranial magneticstimulation. Clin Neurophysiol. 2001;112:720.

19. André-Obadia N, Mertens P, Gueguen A, Peyron R,Garcia-Larrea L. Pain relief by rTMS: Differentialeffect of current flow but no specific action on painsubtypes. Neurology. 2008;71:833-840.

20. Wassermann EM. Risk and safety of repetitivetranscranial magnetic stimulation: Report and sug-gested guidelines from the International Workshopon the Safety of Repetitive Transcranial MagneticStimulation, June 5-7, 1996. Electroencephalogr ClinNeurophysiol. 1998;108:1-16.

21. Rossi S, Hallett M, Rossini PM, Pascual-Leone A;Safety of TMS Consensus Group. Safety, ethical con-siderations, and application guidelines for the use oftranscranial magnetic stimulation in clinical practiceand research. Clin Neurophysiol. 2009;120:2008-2039.

22. Smith MJ, Keel JC, Greenberg BD, et al. Menstrualcycle effects on cortical excitability. Neurology.2002;53:2069-2072.

23. Liebetanz D, Nitsche MA, Tergau F, Paulus W.Pharmacological approach to the mechanisms oftranscranial DC-stimulation-induced after-effects of

672 April 2014

human motor cortex excitability. Brain. 2002;125(Pt10):2238-2247.

24. Coppola G, Pierelli F, Schoenen J. Is the cerebralcortex hyperexcitable or hyperresponsive inmigraine? Cephalalgia. 2007;27:1427-1439.

25. Mulleners WM, Chronicle EP, Palmer JE, KoehlerPJ, Vredeveld JW. Suppression of perception inmigraine: Evidence for reduced inhibition in thevisual cortex. Neurology. 2001;56:178-183.

26. Aurora SK, Barrodale P, Chronicle EP, MullenersWM. Cortical inhibition is reduced in chronic andepisodic migraine and demonstrates a spectrum ofillness. Headache. 2005;45:546-552.

27. Valeriani M, Fierro B, Brighina F. Brain excitabilityin migraine: Hyperexcitability or inhibited inhibi-tion? Pain. 2007;132:219-220, author reply 220-2.

28. Alam Z, Coombes N, Waring RH, Williams AC,Steventon GB. Plasma levels of neuroexcitatoryamino acids in patients with migraine or tensionheadache. J Neurol Sci. 1998;156:102-106.

29. Vaccaro M, Riva C, Tremolizzo L, et al. Platelet glu-tamate uptake and release in migraine with andwithout aura. Cephalalgia. 2007;27:35-40.

30. Siniatchkin M, Sendacki M, Moeller F, et al. Abnor-mal changes of synaptic excitability in migraine withaura. Cereb Cortex. 2011;22:2207-2216.

31. Coppola G, Pierelli F, Schoenen J. Habituation andmigraine. Neurobiol Learn Mem. 2009;92:249-259.

32. Pierelli F, Iacovelli E, Bracaglia M, Serrao M,Coppola G. Abnormal sensorimotor plasticity inmigraine without aura patients. Pain. 2013;154:1738-1742.

33. Peurala SH, Müller-Dahlhaus JF, Arai N, ZiemannU. Interference of short-interval intracortical inhibi-tion (SICI) and short-interval intracortical facilita-tion (SICF). Clin Neurophysiol. 2008;119:2291-2297.

34. Shin HC, Chapin JK. Mapping the effects of motorcortex stimulation on single neurons in the dorsalcolumn nuclei in the rat: Direct responses and affer-ent modulation. Brain Res Bull. 1989;22:245-252.

35. Shin HC, Chapin JK. Movement induced modula-tion of afferent transmission to single neurons in theventroposterior thalamus and somatosensory cortexin rat. Exp Brain Res. 1990;81:515-522.

36. Ziemann U, Paulus W, Nitsche MA, et al. Consen-sus: Motor cortex plasticity protocols. Brain Stimul.2008;1:164-182.

37. Di Lazzaro V, Oliviero A, Berardelli A, et al. Directdemonstration of the effects of repetitive trans-

cranial magnetic stimulation on the excitability ofthe human motor cortex. Exp Brain Res. 2002;144:549-553.

38. Inghilleri M, Conte A, Frasca V, Gilio F, LorenzanoC, Berardelli A. Synaptic potentiation induced byrTMS: Effect of lidocaine infusion. Exp Brain Res.2005;163:114-117.

39. Inghilleri M, Gilio F, Conte A, et al. Topiramate andcortical excitability in humans: A study with repeti-tive transcranial magnetic stimulation. Exp BrainRes. 2006;174:667-672.

40. Richards DA. Regulation of exocytic mode in hip-pocampal neurons by intra-bouton calcium concen-tration. J Physiol. 2010;588(Pt 24):4927-4936.

41. Zucker RS, Regehr WG. Short-term synaptic plas-ticity. Annu Rev Physiol. 2002;64:355-405.

42. Frank CA, Kennedy MJ, Goold CP, Marek KW,Davis GW. Mechanisms underlying the rapid induc-tion and sustained expression of synaptic homeosta-sis. Neuron. 2006;52:663-677.

43. Turrigiano G. Homeostatic synaptic plasticity: Localand global mechanisms for stabilizing neuronal func-tion.Cold Spring Harb Perspect Biol. 2012;4:a005736.

44. Moulder KL, Cormier RJ, Shute AA, Zorumski CF,Mennerick S. Homeostatic effects of depolarizationon Ca2+ influx, synaptic signaling, and survival. JNeurosci. 2003;23:1825-1831.

45. Moulder KL, Meeks JP, Mennerick S. Homeostaticregulation of glutamate release in response to depo-larization. Mol Neurobiol. 2006;33:133-153.

46. Wang X, Engisch KL, Li Y, Pinter MJ, Cope TC,Rich MM. Decreased synaptic activity shifts thecalcium dependence of release at the mammalianneuromuscular junction in vivo. J Neurosci. 2004;24:10687-10692.

47. Adams PJ, Rungta RL, Garcia E, van denMaagdenberg AM, MacVicar BA, Snutch TP. Con-tribution of calcium-dependent facilitation to synap-tic plasticity revealed by migraine mutations in theP/Q-type calcium channel. Proc Natl Acad Sci U S A.2010;107:18694-18699.

48. Antal A, Lang N, Boros K, Nitsche M, Siebner HR,Paulus W. Homeostatic metaplasticity of the motorcortex is altered during headache-free intervals inmigraine with aura.Cereb Cortex. 2008;18:2701-2705.

49. Sándor PS, Dydak U, Schoenen J, et al. MR-spectroscopic imaging during visual stimulation insubgroups of migraine with aura. Cephalalgia. 2005;25:507-518.

Headache 673

50. Brighina F, Palermo A, Daniele O, Aloisio A, FierroB. High-frequency transcranial magnetic stimula-tion on motor cortex of patients affected bymigraine with aura:A way to restore normal corticalexcitability? Cephalalgia. 2010;30:46-52.

51. Bohotin V, Fumal A, Vandenheede M, et al. Effectsof repetitive transcranial magnetic stimulation onvisual evoked potentials in migraine. Brain. 2002;125(Pt 4):912-922.

674 April 2014