Altered processing of sensory stimuli in patients with migraine

Altered processing of sensory stimuli in patients with migraineMarina de Tommaso, Anna Ambrosini, Filippo Brighina, Gianluca Coppola, Armando Perrotta, Francesco Pierelli, Giorgio Sandrini, Massimiliano Valeriani, Daniele Marinazzo, Sebastiano Stramaglia and Jean Schoenen Abstract Migraine is a cyclic disorder, in which functional and morphological brain changes fluctuate over time, culminating periodically in an attack. In the migrainous brain, temporal processing of external stimuli and sequential recruitment of neuronal networks are often dysfunctional. These changes reflect complex CNS dysfunction patterns. Assessment of multimodal evoked potentials and nociceptive reflex responses can reveal altered patterns of the brain's electrophysiological activity, thereby aiding our understanding of the pathophysiology of migraine. In this Review, we summarize the most important findings on temporal processing of evoked and reflex responses in migraine. Considering these data, we propose that thalamocortical dysrhythmia may be responsible for the altered synchronicity in migraine. To test this hypothesis in future research, electrophysiological recordings should be combined with neuroimaging studies so that the temporal patterns of sensory processing in patients with migraine can be correlated with the accompanying anatomical and functional changes.
de Tommaso, M. et al. Nat. Rev. Neurol. 10, 144–155 (2014); published online 18 February 2014; Introduction
Migraine is the most prevalent neurological disorder in
peripheral and central portions of the trigeminovascular the general population, with a cumulative lifetime inci­ system—the main pain­signalling system of the brain. dence of 43% in women and 18% in men.1 The episodic The relative importance and exact sequence of activa­ form of migraine is characterized by recurrent head­ tion of these structures during a migraine attack might ache attacks, which are often accompanied by nausea, vary with the migraine type, and remains a topic for vomiting photophobia or phonophobia.2 Some patients develop chronic migraine (at least 15 days of head­ The temporal precision and noninvasiveness of ache per month, including at least 8 days with typical electro physiological methods make them particu­ migraine attacks).2 In about 20% of patients, migraine larly well­suited to study of the cyclic functional brain attacks are preceded by or associated with an aura com­ changes associated with migraine.5 Investigators using posed of transient focal neurological symptoms, such as these techniques have demonstrated that the migrainous scintillating scotomata (blurred areas in the visual field), brain has altered functioning between migraine attacks, University of Bari, Italy par aesthesias or language disturbances. As interictal and that this brain dysfunction undergoes cyclic changes (M.d.T., S.S.). University symptoms and overt brain lesions are absent, migraine up to initiation of the attack.6 Various electrophysiologi­ of Palermo, Italy (F.B.). is commonly considered to be a prototypic functional cal parameters have been studied in migraine research, G. B. Bietti Foundation, IRCCS, Italy (G.C.). brain disorder.
including multimodal evoked potentials, steady­state C. Mondino Institute of The common migraine types, migraine with and visual evoked responses, noxious evoked cortical Neurology Foundation, IRCCS, Italy (G.S.). without aura, are determined by complex inter­ responses, and nociceptive reflexes. The results have Ospedale Pediatrico actions between multiple additive genetic, environ­ provided three main sets of observations, which were Bambino Gesù, IRCCS, mental, hormonal and endogenous (cognitive and consistent across most studies. First, between attacks, a Italy (M.V.). University of Ghent, Belgium emotional) factors.3 These factors modify dynamic stimulus­frequency­dependent increase occurs in photic (D.M.). Liège University, interactions between various brain areas and compo­ driving and synchronization of EEG alpha (8–13 Hz) and Belgium (J.S.). Headache Clinic, nents that define the individual's level of susceptibil­ beta (13–30 Hz) rhythms. Second, the interictal migrain­ Istituto di Ricovero e ity to migraine. The susceptibility level fluctuates and ous brain is characterized by a habituation (or adapta­ Cura a Carattere at times becomes sufficiently intense to precipitate a tion) deficit of cortical evoked responses to repetitive, Scientifico (IRCCS) Neuromed, migraine attack. The neural components involved in non­noxious sensory stimuli—this deficit normalizes Via Atinense 18, susceptibility to migraine include the cerebral cortex, during an attack. Third, responses or reflexes evoked Pozzilli, 86077 Isernia, Italy (A.A., A.P., F.P.).
brainstem, hypothalamus and thalamus, as well as by noxious stimuli also fail to habituate interictally, but this abnormality does not reverse during an attack. It Correspondence to: A.A. Competing interests should, however, be noted that not all studies confirmed The authors declare no competing interests.
144 MARCH 2014 VOLUME 10
2014 Macmillan Publishers Limited. All rights reserved Key points
Another interesting aspect of visual y induced changes ■ Migraine is the most prevalent neurological disorder in the general population in EEG recordings is that they might differ between and exerts a considerable societal burden; in some patients, migraine becomes migraine types. Some SSVEP studies found no differ­ unremittingly chronic ences between migraine with and without aura,13 but ■ Electrophysiological studies can characterize the abnormal functioning of the one study showed that interhemispheric SSVEP asym­ migrainous brain between, immediately before and during attacks, and aid metry was increased in about half of patients affected monitoring of the effects of therapeutic interventions by migraine with aura, whereas in those with migraine ■ Most electrophysiological studies of migraine describe functional changes without aura, the amplitude of the second harmonic was between attacks, including hyperresponsivity to repeated sensory stimuli increased.14 Another group found an increased ampli­ with abnormal temporal processing, malfunctioning sequential recruitment of neuronal networks, and impaired habituation tude of the second harmonic in both migraine groups, ■ The abnormalities of sensory processing vary over the migraine cycle: they but an augmented amplitude of the fourth harmonic at worsen preictally but tend to disappear during the attack; furthermore, the high spatial frequency only in patients who had migraine abnormalities differ between episodic and chronic migraine with aura.15 The investigators interpreted their results as ■ Refined neurophysiological investigations suggest that the cyclic brain reflecting increased responsivity of the primary visual dysfunctions in migraine might be related to an abnormal cross-talk between cortex in both types of migraine, albeit with extension thalamus and cortex (thalamocortical dysrhythmia) of this increased responsivity to include secondary ■ Understanding the dysfunction of temporal information processing in visual areas in migraine with aura. This hypothesis is migraine paves the way for novel acute and preventive therapies, including pathophysiology-based neuromodulatory techniques being further tested in studies of EEG mapping during intermittent visual stimulation, as described below.
In this Review, we describe the data on alterations of neuronal processing in patients with migraine, affect­ The role of neuronal networks in determining responsiv­ ing habituation, potentiation, summation, sequential ity of the brain to visual stimulation can be assessed com­ dipolar source activation, and synchronization. We prehensively by studying the synchronization and causal provide an overview of these neurophysiological studies connections of different brain areas using nonlinear and describe the novel methods used to explore func­ analysis methods. In healthy individuals, the oscillatory tional brain connectivity in the migrainous brain. We activity in the alpha range is suppressed during flicker pay particular attention to the temporal dimension of stimulation, possibly as a result of desynchronization.16 these abnormalities, which seems crucial to understand­ By contrast, in patients who have migraine without aura, ing the functional brain changes in migraine and their the alpha rhythm remains highly synchronized across different brain areas during visual stimulation.17 This pattern does not depend on the alpha amplitude, but EEG changes induced by visual stimuli
pertains to the synchrony of temporal activation and Increased photic driving dynamic interactions (such as functional connectiv­ Many studies have focused on steady­state visual evoked ity) between brain areas,18 and to their modification by potentials (SSVEPs), which are the EEG response to sensory stimuli. Some researchers have suggested that repetitive visual stimulation. SSVEPs are not generated functional connectivity is determined by both cortico­ by amplitude modulation; instead, they primarily result cortical and thalamocortical loops.19 The mechanisms from phase alignment of the ongoing background EEG underlying temporal synchrony of EEG rhythms are activity7 with the changes in frequency of the repeti­ not simply a result of the balance between excitatory tive stimulus. This phenomenon, called photic driving, and inhibitory inputs. Anticonvulsants, for instance, reflects the tendency of cortical neurons to synchronize modulate alpha rhythm synchronization differentially: their firing with the frequency of the visual stimuli.
topiramate, an established migraine­ preventing drug, has Although normal brain activity is entrained by repeti­ no effect on alpha oscillations, whereas levetiracetam, tive low­frequency (±10 Hz) light stimuli, increased which might also be effective in migraine prevention, photic driving has been described in response to reduces alpha synchronization.20 Low­frequency repeti­ medium­frequency (±20 Hz) light stimuli in patients tive transcranial magnetic stimulation (rTMS), which with migraine, and is called the H response.8 SSVEPs to has been studied as a preventive treatment in migraine flash stimuli in the medium­frequency range confirmed and is thought to inhibit the underlying occipital cortex, increased photic driving in individuals with migraine, has no effect on alpha phase synchronization.21 without any relation to migraine severity or duration.9,10 Oscillatory properties of neuronal networks can be This observation was interpreted as hyper­responsivity accurately assessed by measuring the directed flow of of the brain to visual stimuli. Further analysis showed information between their components, using Granger that in patients with migraine, SSVEP amplitude was causality22,23 or dynamic causal modelling,24 both of less stable over time than in controls.11 Fluctuation of which measure effective connectivity. Granger causal­ increased photic drive over the migraine cycle has also ity detects connectivity only in linear data; however, been reported.10,12 This instability, changes over the a modified version, kernel Granger causality,23can migraine cycle, and methodological differences prob­ be used to infer direct dynamic influences from non­ ably explain some of the contradictory results reported linear signals such as EEG data. In a preliminary study, in the literature.12 individuals who had migraine without aura showed NATURE REVIEWS NEUROLOGY VOLUME 10 MARCH 2014 145
2014 Macmillan Publishers Limited. All rights reserved properties and characteristics of habituation27 have been revised and refined,28 but the underlying neural mecha­ nisms are still not completely understood. Habituation has multiple roles, ranging from pruning of irrelevant information to protection of the cerebral cortex against overstimulation. This phenomenon has been studied to disentangle the neuronal substrates of behaviour, learn­ ing processes, and processing of CNS information in Migraine without aura (n = 19) health and disease.29–32 Migraine with aura (n = 19) Controls (n = 11) The majority of interictal evoked potential studies in patients with migraine support the notion that the migrainous brain is characterized by impaired habitu­ ation to repetitive stimuli. The habituation deficit is observed across several sensory modalities, and is usual y accompanied by a normal to low amplitude of early responses in averaged data. Lack of habituation is the most prominent (and probably genetically determined) consequence of the functional brain abnormality that characterizes many migraine patients between attacks.33 Of note, the abnormal visual information processing that occurs in migraine between attacks corresponds neither to sensitization nor to dishabituation (that is, restoration to the full strength of a response previously weakened by habituation). It is accompanied by initially decreased or normal amplitude of response after a small number of stimuli, followed by a stable amplitude, or even a transient amplitude increase (potentiation).34–43 Figure 1 Temporal evolution in effective connectivity, as
The first evidence for altered interictal habituation in revealed by kernel Granger causality analysis of averaged patients with migraine came from studies of contingent EEG data. Recordings were taken from patients with negative variation (CNV), a slow event­related corti­ migraine and healthy controls during a 10 s flash cal response representing higher mental functions.44–47 stimulation at a 21 Hz frequency. a In the alpha band,
Subsequently, deficient habituation was demonstrated causal/effective connections across scalp derivations are for another event­related potential, P300, which is eli­ weaker in individuals with migraine without aura than in cited in the process of decision­making after visual48 those with migraine with aura or individuals with no
migraine (healthy controls). b By contrast, individuals
or auditory49,50 stimulation. Deficient habitu ation was with migraine with aura have increased causality across also subsequently described for several other modality­ cortical areas in the beta band. This phenomenon might be specific evoked potentials: pattern­reversal visual subtended by increased cortical activation in migraine with evoked potentials (VEPs),33–42 visual evoked magneto­ aura during visual stimulation. The figure is based on data encephalographic (MEG) responses,43 auditory evoked from a study by de Tommaso et al.25 potentials (AEPs),51,52 and somatosensory evoked poten­ tials (SSEPs).53–55 Several other studies, however, were increased phase synchronization in the alpha band and not able to reproduce these results and found no habitu­ reduced connectivity during intermittent flash stimula­ ation deficit in individuals with migraine, possibly tion, whereas those who had migraine with aura dis­ because of differences in the methods used or selection played clear desynchro nization in the beta frequency of patients.56–62 range and increased connectivity during visual stimula­ The reasons for the discrepant results of habituation tion (Figure 1).25 Given that brain activation is now studies are not fully understood. Insufficient blinding of described in terms of increased connectivity of different the investigator has been suggested as a possible culprit;63 functional brain networks, visual stimulation seems to however, since the same researchers have found the same induce a more vigorous cortical activation and spread result (that is, normal habituation) in individuals with of information in migraine with aura than in migraine migraine in both blinded and nonblinded studies,57 and without aura (which is charac terized by weak interaction lack of habituation has also been reported in a blinded between cortical regions), possibly because of a prevalent study, this factor is unlikely to be the sole cause.64 Factors resonance of rhythmic activity generated at subcortical directly related to the pathophysiology of migraine are and thalamic levels.22 probably involved. For instance, the habituation deficit does not remain constant in individuals with migraine. Evoked responses to non-noxious stimuli
The deficit varies over the migraine cycle: it is profound Impaired habituation during the interictal state, normalizes briefly before Habituation—a response decrement as a result of and during the attack and increases again a few days repeated stimulation26—is a multifactorial process. The after the attack.65 Moreover, genetic variants can have an 146 MARCH 2014 VOLUME 10
2014 Macmillan Publishers Limited. All rights reserved effect on habituation profiles.66,67 Finally, spontaneous contrasts with the interictal habituation deficit observed clinical worsening or improvement of attack frequency in episodic migraine. Interestingly, the habituation can influence the baseline level of thalamocortical activa­ deficit reappears when patients evolve from chronic tion68,69 and, hence, the degree of habituation in patients to episodic migraine.83 Since the response pattern in chronic migraine is indistinguishable from that observed during migraine attacks,43,51–62,70–72 our research group has Variation over the migraine cycle suggested that patients with chronic migraine are locked Episodic migraine in an ictal­like state.84 Episodic migraine is, by definition, a cyclic disorder. The most prevalent factor associated with the tran­ The attack itself is not an abrupt event, but the result sition from episodic to chronic migraine is acute of a sequential process that might start as the so­called medication overuse. In medication overuse headache prodromal or premonitory symptoms several hours (MOH), the cortical response pattern suggests that the before the aura or the headache. Moreover, attack fre­ brain is locked in a preictal state, characterized by an quency varies over the patient's lifetime. It is, therefore, increased amplitude of responses to intermittent stimuli of major pathophysiological interest to study the changes (sensitization) and a consistent deficit of habituation to in brain responsivity associated with various stages of continuous or repetitive stimuli.54 This pattern might the migraine cycle. During the days preceding an attack, vary with the class of drug overused. In triptan over­ CNV and P300 habituation is minimal, and the ampli­ users, the initial SSEP amplitude is normal, whereas it is tude of these responses is maximal.70,71 Within the increased in overusers of NSAIDs and in those overusing 12–24 h immediately preceding an attack, and during drugs from both classes.54,85 In both groups of overusers, the attack, habituation of evoked potentials normalizes. however, SSEP habituation was abnormal.
This pattern has been shown for CNV,36,70,71 VEP,57,72 and SSEP61 amplitudes, and for visual P300 latency.73 The Possible mechanisms of habituation R2 component recorded during blink reflexes—evoked Genetic predisposition is likely to influence the brain's by an electrical stimulus delivered with a classic non­ responsivity patterns, although its effects are variable nociceptive surface electrode—showed a habituation between patients and migraine types. In migrainous deficit in patients before a migraine attack,76 although in child–parent pairs, habituation of evoked potentials another study only slight habituation abnormalities were has a clear familial pattern.66,71 Moreover, in asympto­ found interictally.77 In a longitudinal study of brainstem matic individuals who have a first­degree relative with AEPs, habituation of wave IV–V amplitude was deficient migraine, and are thus at risk of developing migraine in patients with migraine, but did not change over the during their lifetime, cortical evoked potentials79 and migraine cycle.78 nociceptive blink reflexes86 (nBRs, discussed under To our knowledge, no single satisfactory explanation processing of noxious stimuli below) showed amplitude exists for the cyclic nature of episodic migraine, except and habituation abnormalities similar to those found for the one related to the ovarian cycle and variations interictally in people with migraine.
in sex hormone levels. Nonetheless, various experimen­ The neural mechanisms underlying habituation and tal data suggest some interesting avenues for further its impairment in patients with migraine remain poorly research. For instance, cortical responsivity is cyclic in understood.87 In theory, habituation deficits could be due individuals with migraine,71 and varies in parallel with to increased excitatory mechanisms, decreased activity changes in platelet serotonin content.73 The period­ of inhibitory interneurons, or reduced baseline activa­ icity of neurophysiological brain activity might also tion of sensory cortices according to the ‘ceiling' theory. be related to psychophysical, genetic66,79 or metabolic This theory postulates that an individual's sensory cor­ factors,80 or to the biorhythms of hypothalamic activity.81 tices have variable baseline activation levels, but their Migraine periodicity might thus be the result of several maximum activation level (the ceiling) remains constant. inter acting biological cycles. Indeed, the migraine cycle During repetitive stimulation, the maximum activation is probably caused not by a single determining factor, level is reached rapidly, and subsequently the response but by a complex interplay between intrinsic cerebral, amplitude decreases sharply (habituation) in individu­ hormonal and environmental factors acting on a geneti­ als with normal baseline activation, while habituation cally predisposed nervous system. Disentanglement is delayed or absent in individuals in whom baseline of this interplay is a challenge for future research, and activation is low.88 Both increased cortical excitability will be a prerequisite for the development of effective and decreased activity of inhibitory neurons would be expected to give rise to a high initial response ampli­ tude, indicating genuine hyperexcitability, and a linear Chronic migraine decrease on habituation. The ceiling theory can also Cortical responsivity is different in episodic and chronic account for the normal or decreased initial amplitude migraine. For instance, the initial amplitude of visual and the nonlinear and cyclic changes in habituation.
MEG responses (P100m) was greater in chronic migraine Studies of high­frequency oscil ations (HFOs) embed­ than in interictal episodic migraine.82 Moreover, these ded in evoked cortical responses have contributed to our responses show substantial habituation (comparable to understanding of the abnormal information processing that of healthy controls) to repetitive stimuli,82 which in migraine. The amplitude of early HFOs embedded in NATURE REVIEWS NEUROLOGY VOLUME 10 MARCH 2014 147
2014 Macmillan Publishers Limited. All rights reserved Response habituation Interictal lack of Ictal normalization of Normalization of interictal lack in healthy controls habituation in migraine habituation in migraine of habituation after HF-rTMS Decreased Increased Increased Figure 2 Cortical response patterns during the migraine cycle. This schematic overview shows amplitude changes in the
N20–P25 component of averaged EEG recordings in patients with migraine and healthy controls. a HFOs and
b   somatosensory evoked potentials. In healthy controls (panel 1), the N20–P25 component habituates, and early HFOs
(reflecting thalamocortical drive) are greater than late HFOs (generated by intrinsic cortical activation). In patients with
migraine between attacks (panel 2), habituation is absent and early HFOs are reduced, although late HFOs are normal. During
a migraine attack (panel 3), habituation and early HFOs normalize. After 10 Hz HF-rTMS is applied over the somatosensory
cortex in patients with episodic migraine (panel 4), the interictal lack of habituation reverses, and both early and late HFOs
increase. Abbreviations: HFO, high-frequency oscillation; HF-rTMS, high-frequency repetitive transcranial magnetic stimulation.
the common SSEPs, which is thought to reflect spiking of P50 middle­latency AEPs92 and the significant habit­ activity in thalamocortical cholinergic afferents, is uation deficit in late visual­evoked high­frequency decreased interictally in patients with migraine and activity (oscillations in the gamma range, 20–35 Hz),93 normalizes during attacks, whereas that of late HFOs, in comparison with healthy controls. Taken together, which probably reflect the activity of inhibitory cor­ these studies indicate dysfunction of thalamocorti­ tical interneurons, remains normal89 or decreases90 cal oscillatory networks, and patients with migraine between attacks. Moreover, a reduction in amplitude might, therefore, be considered to have thalamocortical of early HFOs is associated with worsening of the clini­ cal course of migraine.68 Contrasting with these results, The thalamocortical dysrhythmia theory postulates increased amplitudes of early and late HFOs has been that when anatomical or functional disconnection of the reported between attacks in patients with migraine.91 thalamus from subcortical areas is present, the rhythmic These dis parate findings might be a result of differences thalamocortical activity might change to favour low­ in recording parameters and patient selection.
frequency activity (mainly 4–7 Hz theta waves). At the In patients with migraine, activation of the sensori­ cortical level, this change will result in reduced firing motor cortex induced by 10 Hz rTMS increased the rates of excitatory pyramidal cells at the beginning of amplitude of early and late HFOs in SSEPs, and induced stimulation, and of fast­spiking inhibitory interneurons habituation of the broadband SSEP.55 rTMS significantly during stimulus repetition.94,95 Reduced firing of fast­ increased the amplitude of late HFOs, but had no effect spiking interneurons leads to disinhibition of adjacent on either early HFOs or habituation of the broadband cortical columns, which is reflected by a progressive rise SSEP in nonmigrainous controls, probably because in high­frequency gamma band oscillations—the so­ their thalamocortical activity was already maximal at called edge effect.95 This theory could explain both the baseline.55 These observations support the hypothesis reduced thalamic and thalamocortical activity observed that the habituation deficit in patients with migraine is with HFOs, and the rise in late visual­evoked gamma due to reduced thalamic activation and, hence, reduced band oscillations. baseline activation of sensory cortices. Concordant data Several findings support the thalamocortical dysrhyth­ indicate that the interictal habituation deficit and low mia theory. Short­range lateral inhibition in the visual initial amplitude of VEPs in individuals with migraine cortex is more pronounced in migraine patients than normalizes after 10 Hz rTMS over the visual cortex.36 in healthy volunteers at the beginning of the stimulus Further evidence that control of thalamocortical activ­ session.65 Moreover, short­range lateral inhibition in ity is abnormal in people with migraine between attacks the visual cortex increases over successive responses is suggested by the marked reduction in sensory gating in people with migraine, but remains unchanged in 148 MARCH 2014 VOLUME 10
2014 Macmillan Publishers Limited. All rights reserved healthy controls.65 Several quantitative EEG studies in Processing of noxious stimuli
individuals with migraine have shown a widespread Another feature that is present in patients with increase in slow (mostly theta) oscillatory activity, migraine concerns the altered processing of nociceptive chiefly over temporo­occipital areas,96,97 which similarly stimuli, which has been studied using nocicep­ concords with the thalamocortical dysrhythmia theory.
tive trigeminal and biceps femoris reflexes, as well as thermonociceptive­induced cortical evoked responses.
Amplitude–stimulus intensity functionAnother time­related modality of stimulus processing The nociceptive flexion reflex that seems to be altered in people with migraine is the Pain disorders are commonly accompanied by central progressive amplitude adaptation of cortical responses sensitization, which amplifies the CNS response to to repetitive stimuli of increasing intensity, which is painful stimuli. This amplification also occurs during referred to as the amplitude–stimulus intensity function. migraine attacks108 and worsens with increasing attack When stimuli are delivered at increasing intensity, the frequency.109 One mechanism underlying central sensi­ evoked cortical responses increase in certain individu­ tization is the activity­dependent change in excitability als, but decrease in others.98 This so­called augmenting– of central nociceptors, which results in abnormal ampli­ reducing response has been widely studied, mainly in fication of pain sensation in physiological nociception— the context of auditory stimuli. The intensity depend­ a phenomenon referred to as temporal summation of pain ence of AEPs (IDAP) is expressed by the amplitude– stimuli110 that is equivalent to ‘wind­up' phenomenon stimulus intensity slope of the cortical N1–P2 wave, (facilitation of wide­dynamic­range nociceptive neurons where N1 is the greatest negative component 60–150 ms located in the deep laminae of the spinal cord dorsal horn post­ stimulus and P1 is the greatest positivity from and the spinal trigeminal nucleus after constant­intensity 120–200 ms. Interestingly, IDAP correlates inversely with stimulation of C fibres) in animal experiments.111 central serotonergic transmission, as evaluated indirectly The nociceptive flexion or withdrawal reflex (NWR) is by biochemical and pharmacological methods.99 a reliable measure of spinal nociception, as demonstrated Although the grand average of long­latency AEPs by the fact that it requires Aδ­fibre activation, that the has normal latency and amplitude in patients with reflex threshold is related to the pain perception thresh­ migraine,57,100 IDAP is significantly increased inter­ old, and that the reflex magnitude correlates positively ictally compared with healthy volunteers in most51,52,72— with pain intensity ratings.112,113 Temporal summation although not all57—studies. IDAP normalizes the day of pain develops in paral el with tem poral summation of before and during the migraine attack, similarly to VEP the NWR of the lower limbs, reflected by a progres­ habituation.72 In fact, the interictal increase in IDAP sive increase in magnitude of the NWR after constant­ in people with migraine can be attributed to a habitu­ intensity electrical stimulation (which activates both ation deficit of the cortical response to high­intensity Aδ and C fibres,113–115 and is inhibited by N­methyl­d­ auditory stimuli.52 IDAP is also strongly influenced by aspartate receptor antagonists).116 Interestingly, descend­ sensory overload.101 Indeed, when IDAP is assessed ing pain control systems modulate temporal summation during intense flash stimulation, two subgroups of of the NWR,117 and might be dysfunctional in a number of patients with migraine can be identified—one reacts chronic pain dis orders, including migraine. For example, to the stimulus by a decrease in IDAP, as do controls, studies of tem poral summation of the NWR in people whereas the other reacts by an increase in IDAP. The with migraine show facilitation of temporal pain pro­ underlying neuro biological basis of this difference cessing between attacks.118 Administration of a nitric between clinical y similar patients is unknown, but might oxide donor, such as glyceryl trinitrate, induces transi­ be related to differences in genetic background and/or tory facilitation of temporal summation of the NWR within 120 min in patients who will go on to develop a An increased IDAP (that is, an augmenting pattern) glyceryl trinitrate­triggered migraine attack several hours suggests the presence of decreased central serotonergic transmission.102,103 A high IDAP correlates positively with In individuals with chronic headache, such as MOH clinical symptoms of major depression104 that are thought in the setting of episodic migraine, the threshold for to be associated with decreased serotonergic signal­ temporal summation of the NWR is markedly reduced ling, and normalizes in depressed patients treated with compared with that in controls or in patients with epi­ selective serotonin reuptake inhibitors.105 IDAP abnor­ sodic migraine, which indicates strong facilitation in malities also correlate with personality traits thought the temporal processing of pain.118 In patients with to be associated with decreased serotonergic transmis­ MOH, the pain­suppressing effect of supraspinal diffuse sion in individuals with migraine.106 Treatment with noxious inhibitory control (pain inhibition by hetero­ migraine­preventing drugs such as β­blockers, which topic painful stimulation) on temporal summation of increase serotonergic transmission, normalizes the the NWR is deficient.118 This effect, which in humans increased interictal IDAP in patients with migraine.107 is termed conditioned pain modulation,120 can be tested All things considered, the increased IDAP in migraine by heterotopic application of a painful cold stimulus.119 could be secondary to reduced activity of raphe corti­ The deficits in conditioned pain modulation or supra­ cal mono aminergic pathways, which causes low baseline spinal diffuse noxious inhibitory control and facilitation activation levels of auditory cortices.
of temporal summation of the NWR normalize after NATURE REVIEWS NEUROLOGY VOLUME 10 MARCH 2014 149
2014 Macmillan Publishers Limited. All rights reserved neuropathic pain in healthy volunteers,127 and with the caudal displacement of cortical evoked potentials in the cingulate gyrus after intramuscular nociceptive stimulation of the trapezius muscle in patients with migraine.128 This difference with the data on LEPs can be explained by the different methodologies used, which involved stimulation of different nociceptive afferents.129 Similarly to the cortical evoked potentials elicited by non­noxious stimuli, LEPs75,130 and CHEPs131 show habituation deficits in patients with migraine between attacks. However, in contrast with non­noxious cortical evoked potentials, the lack of habituation of LEPs persists during the attack, and is associated with an increased Patients with migraine (GTN) N2–P2 amplitude.132 Patients with migraine (placebo) Nonlinear analysis of ongoing EEG changes reveals Controls (placebo) subtle changes in the cortical response to painful laser stimuli in patients with migraine.133 For example, indi­ viduals with episodic migraine have markedly reduced predictability of their EEG rhythms after the laser stimu­ Temporal summation threshold of the nociceptive withdra lus compared with healthy individuals, although their Figure 3 Facilitation of temporal pain processing between migraine attacks.
averaged LEPs seem normal; however, the averaging Facilitation of the temporal summation threshold of the biceps femoris nociceptive technique used to extract evoked potentials from the withdrawal reflex is markedly more facilitated by glyceryl trinitrate administration background EEG signals might neglect subtle changes (versus placebo) in patients with migraine than in healthy controls. Abbreviation: in the processing of pain by the brain. Future studies GTN, glyceryl trinitrate. Permission obtained from John Wiley and Sons, West using analysis of single (nonaveraged) nociceptive Sussex (UK), Eur. J. Pain 15, 482–490 (2011).
evoked potentials, refined neurophysiological techniques and a combination of neurophysiological and imaging drug withdrawal, which could be related to the reduc­ methods will help to characterize the pathophysiologi­ tion in activity of anandamide hydrolase (also known cal features of central processing of pain in patients with as fatty acid amide hydrolase) and, hence, slowing of the migraine. In chronic migraine and MOH, pain­related degradation of endocannabinoids.121 cortical potentials as a response to electrical forehead or forearm stimulation were facilitated, but no change was Nociceptive trigeminal evoked responses observed in the nBR.134 The nBR is evoked in orbicularis oculi muscles by stim­ ulating the supraorbital nerve via a concentric high­ Thalamocortical dysrhythmia in migraine
density electrode, which mainly activates Aδ afferents. Given the results of the above­described neurophysio­ This reflex is mediated via interneurons of the spinal logical studies, the pathogenesis of migraine seems to be trigeminal nucleus. Migraine is characterized by an driven by complex dysfunction of thalamocortical con­ interictal deficit of nBR habituation during both short74 nectivity and temporal activation of neuronal networks. and long38 series of stimuli. nBR habituation normal­ Thalamocortical dysrhythmia might also explain the izes during migraine attacks,74 and individuals at risk phenomena observed in patients with migraine who are of developing migraine lack nBR habituation deficits,86 treated with transcranial neuromodulation techniques— whereas habituation of nociceptive laser­evoked poten­ for instance, the increased variability of dynamic changes tials (LEPs, discussed further below) remains deficient.75 in excitability135 or the paradoxical homeostatic cer­ Patients with migraine also show temporal summation ebral plasticity.136–138 In a proof­of­concept study, the plastic cortical changes induced by rTMS were found Brief radiant heat pulses generated by CO laser stim­ to be inversely related to thalamocortical activation.139 ulation or contact thermode­delivered stimuli excite This observation suggests that the paradoxical effects Aδ and C fibre thermonociceptors in superficial skin observed after rTMS in patients with migraine might be layers.123 In turn, the Aδ fibre input generates LEPs or a consequence of abnormal thalamocortical drive, which contact­heat evoked potentials (CHEPs) in the cortex. impairs short­term and long­term changes in cortical The N2–P2 component of LEPs and CHEPs is thought synaptic effectiveness, and finally leads to maladaptive to be generated in the posterior part of the anterior responses. Taken together, the dysfunctions found in the cingulate cortex and in the bilateral insula.124 migrainous brain suggest an impairment of thalamo­ Compared with healthy controls and people with epi­ cortical control of temporal activation of different sodic migraine between attacks, the brain distribution of LEP is shifted rostrally in patients with migraine during Thalamocortical dysrythmia has also been pro­ an attack125 and in patients with chronic migraine.126 posed to underlie other functional brain disorders.140,141 This anterior shift of activation contrasts with the pos­ In patients with chronic neuropathic pain, over­ terior shift of LEPs observed during capsaicin­induced activation in the theta and beta range—suggestive of 150 MARCH 2014 VOLUME 10
2014 Macmillan Publishers Limited. All rights reserved trigeminovascular system and induce a migraine attack. Mitochondrial ATP Also, whether the abnormal temporal processing of noci­ of sensory cortices Deficient habituation ceptive information predisposes to migraine attacks, of evoked potentials central sensitization and, possibly, chronic migraine is not proven, but would intuitively seem probable. However, and connectivity the fact that the interictal cortical hyperrespons ivity to Amygdala, hypothalamus sensory stimuli in migraine can be alleviated by neuro­ PAG, monoaminergic nuclei stimulation techniques55 (see below) and by preventive (familiar or sporadic antimigraine drugs, both of which also decrease attack hemiplegic migrane) frequency,107 supports indirectly the hypothesis that the Central sensitizing brain dysfunction between attacks could predispose of pain processing patients to recurrent attacks. Considering that the cer­ Potentials evoked by Cortical spreading ebral energy reserve (ATP content) between attacks is sig­ Nociceptive blink reflex nificantly lower in individuals with migraine compared Temporal summation of nociceptive withdrawal reflex with healthy individuals,80 it is tempting to speculate that the cortical hyperresponsivity might contribute to dis­ ruption of the brain's metabolic homeostasis by increas­ Amplifies and persists in ing energy demand, thereby initiating the biochemical cascade that leads to the migraine attack.6 Figure 4 A neurophysiological model of migraine pathogenesis. Activation of the
trigeminovascular system—the main pain-signalling system in the brain—triggers
Prospects for clinical research
migraine headache (1). The migraine aura is caused by CSD, which may or may not The results of MRI studies suggest that migraine is activate the trigeminovascular system. Genetic channelopathies (2) predispose to CSD in the rare familial and sporadic hemiplegic forms of migraine. Interictal associ ated with altered interictal functional connectivity thalamocortical dysrhythmia causes hyperresponsivity of sensory cortices (3) as in subcortical and cortical areas that are devoted to cog­ well as abnormal pain processing (4). The thalamocortical dysrhythmia itself may nitive functions and pain processing.144,145 Connectivity be induced by decreased control from brainstem monoaminergic nuclei (5). was stronger between the periaqueductal grey and Cortical hyperresponsivity combined with a decreased mitochondrial energy several brain areas associated with pain processing, such reserve favours metabolic strain (6). This process could trigger CSD in the cortex as the prefrontal cortex, anterior cingulate and amygdala, and, via subcortical chemosensitive structures, activate the trigeminovascular areas that are very similar to brain regions implicated system. The migraine attack is associated with sensitization of central nociceptive in neurophysiological data on sequential cortical activa­ pathways (7). Abnormalities of responses evoked by noxious stimuli amplify and persist in chronic migraine (8). Dashed lines indicate hypothetical connections, for tion during painful stimuli.125,126,144,146 Diffusion­weighted which there is currently little or no experimental evidence. Abbreviations: CSD, MRI studies showed that microstructural alterations of cortical spreading depression; PAG, periaqueductal grey.
white matter, and thus of functional connectivity, are present across the orbitofrontal cortex, insula, thala­ thalamocortical dysrhythmia—was found in the corti­ mus and dorsal midbrain.147 These alterations might cal ‘pain matrix'. In six patients who were successfully reflect maladaptive plastic changes driven by dysrupted treated by central lateral thalamotomy, the overactivity exo genous and endogeneous multimodal task process­ attenuated along with the pain.142 ing.147 In another fMRI study, thalamic sensitization cor­ Further assessment with modern neuroimaging related with widespread mechanical al odynia during the methods will be required to disentangle the anatom­ migraine attack.148 Moreover, in a diffusion tensor MRI ical correlates of thalamocortical dysrhythmia in study, our research group found dynamic ictal and inter­ migraine. A functional MRI (fMRI) study in people ictal microstructural variations in the thalamus that were with migraine revealed a lack of habituation of the related to the time since the last migraine attack, and blood oxygen level­dependent signal during repetitive seemed to mimic the cyclic neurophysiological changes trigeminal nociceptive stimulation in areas of the pain described above.149 matrix (anterior insula and middle cingulate gyrus).143 Collectively, these observations suggest that a search Interestingly, this difference between patients with for optimal methods of influencing the cortical tem­ migraine and healthy controls was not found for olfac­ poral processing of exogenous stimuli that can trigger tory stimuli, which the researchers attributed to the fact a migraine attack, or methods for modulating endo­ that olfaction is not relayed in the thalamus.
genous trigeminal noxious inputs that lead to central Our research group has proposed that hypofunction­ sensitization and eventually chronic headache, could ing serotoninergic projections to the thalamus and cortex result in novel interventions for migraine prevention. The might cause functional disconnection of the thalamus, modes of action of anticonvulsants or antidepressants, as leading to thalamocortical dysrhythmia and reduced well as of other pharmacological or nonpharmacologi­ cortical habituation (Figure 4).40 cal interventions, such as neuromodula tion methods, It has not yet been demonstrated whether the altered should be reconsidered in terms of their ability to nor­ synchronicity and deficient habituation of neuronal malize the complex abnormalities of brain connectivity responses to external stimuli in migraine has a role in and hyperresponsivity found in patients with migraine.150 the cortical predisposition to spreading depression, For example, noninvasive cortical neuromodulation or in other phenomena that are able to activate the techniques such as rTMS and transcranial direct current NATURE REVIEWS NEUROLOGY VOLUME 10 MARCH 2014 151
2014 Macmillan Publishers Limited. All rights reserved Box 1 Neurophysiological findings associated with migraine
differs between patients with migraine and healthy con­ trols (Box 1). The neuronal networks involved in sensory Several abnormalities of sensory processing may be observed in patients with processing are characterized by different modalities of sequential recruitment under different environmental Interictal light-induced EEG changes or endogenous conditions. The patterns of temporal ■ Stimulus-frequency-dependent increase in photic drive activation have been analysed over a range of neuronal ■ Alpha frequency synchronization and decrease in functional connectivity (in migraine without aura) activities, from progressive changes of neuronal recruit­ ■ Beta frequency desynchronization and increase in functional connectivity ment in the habituation or intensity dependence phenom­ (in migraine with aura) ena, to facilitation of noxious stimuli summation, and Interictal (and ictal) changes in non-noxious sensory evoked potentials complex patterns of variability, phase synchronization ■ Trend towards lower amplitude of averaged responses to brief sequences and causality that are adapted to describe the properties of repeated stimuli of a chaotic and nonlinear system. These intricate pro­ ■ Deficient habituation during prolonged stimulus repetition (normalizes cesses not only differ between patients with migraine and healthy individuals, but also vary according to the phases ■ Increased intensity-dependence of auditory evoked potentials (normalizes of the migraine cycle in the same patient.
The mechanisms underpinning these complex changes Interictal (and ictal) changes in noxious sensory evoked responses are far from being understood, and how they fit into the ■ Deficient habituation of cortical evoked responses (persists during attack) puzzle of migraine pathogenesis is still unclear. Owing ■ Deficient habituation of nociceptive blink reflexes (normalizes during attack)■ Facilitation of temporal summation of the biceps femoris flexion reflex to their complexity, however, the brain dysfunctions are unlikely to be explained simply by an imbalance between Changes in chronic migraine■ Increased amplitude of averaged cortical responses to small numbers of excitatory and inhibitory circuits.87 We propose that repeated non-noxious and noxious stimuli thalamocortical dysrhythmia could be the culprit for ■ Deficient habituation of sensory evoked responses to noxious stimuli despite abnormal central processing of non­noxious and noxious normal habituation of sensory responses evoked by non-noxious stimuli sensory stimuli in patients with migraine, and that the thalamocortical dysrhythmia itself might be caused by genetically determined inadequate control of the stimulation (tDCS) have already been assessed in clini­ thalamus and cortex by monoaminergic (serotonergic) cal trials. Several studies have investigated the hypoth­ projections originating in the brainstem (Figure 4). We esis that the cortex in patients with migraine patients further postulate that the cortical hyperresponsivity to is hyperexcitable between attacks. However, inhibitory sensory stimuli might contribute causally to migraine low­frequency rTMS over the vertex had no superior attack repetition, because it favours excessive energy therapeutic effect to sham stimulation,151 and cathodal expenditure in a brain with a reduced energy reserve.
(inhibitory) tDCS over the occipital cortex had no signifi­ To reduce discrepancies between studies, more atten­ cant preventive effect on migraine attacks , although the tion should be paid to blinding of investigators, so that latter intervention did reduce attack intensity compared accurate clinical data and headache diaries can be col­ with placebo.152 By contrast, in a pilot trial designed to lected before and during testing. In addition, prospec­ assess an alternative hypothesis—that the visual cortex tive studies should be conducted to monitor patients' is not hyperexcitable per se, but, rather, insufficiently clinical fluctuations throughout the migraine cycle. It activated at baseline (as described above87)—anodal will also be of utmost importance to gather more data (facilitatory) tDCS over the occipital cortex significantly on the (neurophysiological) phenotype–genotype cor­ decreased attack frequency and intensity when used as relations in patients with the various migraine types. preventive therapy in patients with migraine.150 Finally, improved insight into the nature of the inter­ The challenge for future research, therefore, lies in ictal dysfunction of temporal information processing in identification of the precise anatomical structures and individuals with migraine will, we hope, pave the way functional networks involved in migraine, and determi­ for novel therapeutic targets, and could herald improved nation of which pharmacological and nonpharmacologi­ cal interventions can optimal y modulate the function of these areas and thereby improve temporal information processing. Such investigations wil involve simultaneous recordings of the above­reported phenomena via neuro­ We initially searched the PubMed database to identify physiological and functional neuroimaging techniques, articles published up to June 2013. The search terms along with the application of nonlinear algorithms to used were "migraine", "electroencephalography", model brain complexity.153 Novel therapeutic inter­ "EEG", "evoked potentials", "habituation", "temporal summation", "nociceptive withdrawal reflex" and "blink ventions can then be tested for their capacity to normal­ reflex", alone and in combination. The literature search ize the anatomical and functional changes associated was updated using the additional keywords "migraine", with migraine and its subtypes.
"habituation" and "evoked potentials" to identify full-text papers written in English and published in peer-reviewed journals up to December 2013, using the PubMed and Most of the data described here suggest that the cortical Google Scholar databases. Reviews were considered only processing of non­noxious and noxious sensory stimuli when they introduced new concepts or hypotheses.
152 MARCH 2014 VOLUME 10
2014 Macmillan Publishers Limited. All rights reserved 1. Stewart, W. F. et al. Cumulative lifetime migraine 20. de Tommaso, M. et al. Effects of levetiracetam (rTMS) in healthy volunteers and migraine incidence in women and men. Cephalalgia 28, vs topiramate and placebo on visually evoked patients. Cephalalgia 26, 143–149 (2006).
1170–1178 (2008).
phase synchronization changes of alpha rhythm 40. Coppola, G. et al. Interictal abnormalities of 2. Headache Classification Subcommittee of the in migraine. Clin. Neurophysiol. 118, 2297–2304 gamma band activity in visual evoked responses International Headache Society. The in migraine: an indication of thalamocortical international classification of headache 21. de Tommaso, M. et al. Lack of effects of low dysrhythmia? Cephalalgia 27, 1360–1367 disorders, 3rd edn. Cephalalgia 33, 629–808 frequency repetitive transcranial magnetic stimulation on alpha rhythm phase 41. Coppola, G. et al. Changes in visual-evoked 3. Goadsby, P. J., Charbit, A. R., Andreou, A. P., synchronization in migraine patients. Neurosci. potential habituation induced by hyperventilation Akerman, S. & Holland, P. R. Neurobiology of Lett. 20, 143–147 (2011).
in migraine. J. Headache Pain 11, 497–503 migraine. Neuroscience 161, 327–341 (2009).
22. Granger, C. W. Investigating causal relations by 4. Bernstein, C. & Burstein, R. Sensitization of the econometric models and crossspectral methods. 42. Coppola, G., Crémers, J., Gérard, P., Pierelli, F. trigeminovascular pathway: perspective and Econometrica 37, 424–438 (1969). & Schoenen, J. Effects of light deprivation on implications to migraine pathophysiology. J. Clin. 23. Marinazzo, D., Pellicoro, M. & Stramaglia, S. visual evoked potentials in migraine without Neurol. 8, 89–99 (2012).
Kernel method for nonlinear Granger causality. aura. BMC Neurol. 11, 91 (2011).
5. Magis, D. et al. Evaluation and proposal for Phys. Rev. Lett. 11, 144103 (2008).
43. Chen, W. et al. Peri-ictal normalization of visual optimalization of neurophysiological tests in 24. Friston, K. J., Harrison, L. & Penny, W. Dynamic cortex excitability in migraine: an MEG study. migraine: part 1--electrophysiological tests. causal modeling. NeuroImage 19, 1273–1302 Cephalalgia 29, 1202–1211 (2009).
Cephalalgia 27, 1323–1338 (2007).
44. Maertens de Noordhout, A., Timsit-Berthier, M., 6. Ambrosini, A., Magis, D. & Schoenen, J. 25. de Tommaso, M., Stramaglia, S., Marinazzo, D., Timsit, M. & Schoenen, J. Contingent negative Migraine—clinical neurophysiology. Handb. Clin. Trotta, G. & Pellicoro, M. Functional and effective variation in headache. Ann. Neurol. 19, 78–80 Neurol. 97, 275–293 (2010).
connectivity in EEG alpha and beta bands during 7. Moratti, S., Clementz, B. A., Gao, Y., Ortiz, T. intermittent flash stimulation in migraine with 45. Kropp, P. & Gerber, W. D. Contingent negative & Keil, A. Neural mechanisms of evoked and without aura. Cephalalgia 33, 938–947 variation during migraine attack and interval: oscillations: stability and interaction with evidence for normalization of slow cortical transient events. Hum. Brain Mapp. 28, 26. Harris, J. Habituatory response decrement in potentials during the attack. Cephalalgia 15, 1318–1333 (2007).
the intact organism. Psychol. Bull. 40, 385–422 123–128 (1995).
8. Golla, F. L. & Winter, A. L. Analysis of cerebral 46. Schoenen, J. & Timsit-Berthier, M. Contingent responses to flicker in patients complaining of 27. Thompson, R. & Spencer, W. Habituation: negative variation: methods and potential interest episodic headache. Electroencephalogr. Clin. a model phenomenon for the study of neuronal in headache. Cephalalgia 13, 28–32 (1993).
Neurophysiol. 11, 539–549 (1959).
substrates of behavior. Psychol. Rev. 73, 16–43 47. Kropp, P. & Gerber, W. D. Contingent negative 9. Puca, F. M., de Tommaso, M., Tota, P. & variation during migraine attack and interval: Sciruicchio, V. Photic driving in migraine: 28. Rankin, C. H. et al. Habituation revisited: evidence for normalization of slow cortical correlations with clinical features. Cephalalgia an updated and revised description of the potentials during the attack. Cephalalgia 15, 16, 246–250 (1996).
behavioral characteristics of habituation. 123–128 (1995).
10. de Tommaso, M. et al. EEG spectral analysis in Neurobiol. Learn. Mem. 92, 135–138 (2009).
48. Evers, S., Bauer, B., Suhr, B., Husstedt, I. W. migraine without aura attacks. Cephalalgia 118, 29. Walpurger, V., Hebing-Lennartz, G., Denecke, H. & Grotemeyer, K. H. Cognitive processing in 324–328 (1998).
& Pietrowsky, R. Habituation deficit in auditory primary headache: a study on event-related 11. de Tommaso, M. et al. Visually evoked phase event-related potentials in tinnitus complainers. potentials. Neurology 48, 108–113 (1997).
synchronization changes of alpha rhythm in Hear. Res. 181, 57–64 (2003).
49. Wang, W. & Schoenen, J. Interictal potentiation migraine: correlations with clinical features. 30. Schestatsky, P. et al. Neurophysiologic study of of passive "oddball" auditory event-related Int. J. Psychophysiol. 57, 203–210 (2005).
central pain in patients with Parkinson disease. potentials in migraine. Cephalalgia 18, 261–265 12. Bjørk, M., Hagen, K., Stovner, L. J. & Sand, T. Neurology 69, 2162–2169 (2007).
Photic EEG-driving responses related to ictal 31. Halberstadt, A. L. & Geyer, M. A. Habituation 50. Siniatchkin, M., Kropp, P. & Gerber, W. D. What phases and trigger sensitivity in migraine: and sensitization of acoustic startle: opposite kind of habituation is impaired in migraine a longitudinal, controlled study. Cephalalgia 31, influences of dopamine D1 and D2-family patients? Cephalalgia 23, 511–518 (2003).
444–455 (2011).
receptors. Neurobiol. Learn. Mem. 92, 243–248 51. Wang, W., Timsit-Berthier, M. & Schoenen, J. 13. Genco, S., de Tommaso, M., Prudenzano, A. M., Intensity dependence of auditory evoked Savarese, M. & Puca, F. M. EEG features in 32. de Tommaso, M. et al. Laser-evoked potentials potentials is pronounced in migraine: an juvenile migraine: topographic analysis of habituation in fibromyalgia. J. Pain 12, 116–124 indication of cortical potentiation and low spontaneous and visual evoked brain electrical serotonergic neurotransmission? Neurology 46, activity: a comparison with adult migraine. 33. Schoenen, J., Wang, W., Albert, A. & Delwaide, P. 1404–1409 (1996).
Cephalalgia 14, 41–46 (1994).
Potentiation instead of habituation characterizes 52. Ambrosini, A., Rossi, P., De Pasqua, V., Pierelli, F. 14. Nyrke, T., Kangasniemi, P. & Lang, A. H. visual evoked potentials in migraine patients & Schoenen, J. Lack of habituation causes high Difference of steady-state visual evoked between attacks. Eur. J. Neurol. 2, 115–122 intensity dependence of auditory evoked cortical potentials in classic and common migraine. potentials in migraine. Brain 126, 2009–2015 Electroencephalogr. Clin. Neurophysiol. 73, 34. Afra, J., Cecchini, A. P., De Pasqua, V., Albert, A. 285–294 (1989).
& Schoenen, J. Visual evoked potentials during 53. Ozkul, Y. & Uckardes, A. Median nerve 15. Shibata, K., Yamane, K., Otuka, K. & Iwata, M. long periods of pattern-reversal stimulation in somatosensory evoked potentials in migraine. Abnormal visual processing in migraine with migraine. Brain 121, 233–241 (1998).
Eur. J. Neurol. 9, 227–232 (2002).
aura: a study of steady-state visual evoked 35. Wang, W., Wang, G. P., Ding, X. L. & Wang, Y. H. 54. Coppola, G. et al. Abnormal cortical responses potentials. J. Neurol. Sci. 271, 119–126 (2008).
Personality and response to repeated visual to somatosensory stimulation in medication- 16. Birca, A., Carmant, L., Lortie, A. & Lassonde, M. stimulation in migraine and tension-type overuse headache. BMC Neurol. 10, 126 (2010).
Interaction between the flash evoked SSVEPs headaches. Cephalalgia 19, 718–724 (1999).
55. Coppola, G., De Pasqua, V., Pierelli, F. & and the spontaneous EEG activity in children 36. Bohotin, V. et al. Effects of repetitive transcranial Schoenen, J. Effects of repetitive transcranial and adults. Clin. Neurophysiol. 117, 279–288 magnetic stimulation on visual evoked potentials magnetic stimulation on somatosensory evoked in migraine. Brain 125, 912–922 (2002).
potentials and high frequency oscillations in 17. Angelini, L. et al. Steady-state visual evoked 37. Ozkul, Y. & Bozlar, S. Effects of fluoxetine on migraine. Cephalalgia 32, 700–709 (2012). potentials and phase synchronization in habituation of pattern reversal visually evoked 56. Oelkers, R. et al. Visual evoked potentials in migraine patients. Phys. Rev. Lett. 93, 038103 potentials in migraine prophylaxis. Headache 42, migraine patients: alterations depend on pattern 582–587 (2002).
spatial frequency. Brain 122, 1147–1155 18. Friston, K. Functional and effective connectivity: 38. Di Clemente, L. et al. Nociceptive blink reflex a review. Brain Connectivity 1, 13–36 (2011).
and visual evoked potential habituations are 57. Sand, T. & Vingen, J. V. Visual, long-latency 19. Silberstein, R. B. Steady-state visually evoked correlated in migraine. Headache 45, auditory and brainstem auditory evoked potentials, brain resonances and cognitive 1388–1393 (2005).
potentials in migraine: relation to pattern size, processes. In Neocortical Dynamics and Human 39. Fumal, A. et al. Induction of long-lasting changes stimulus intensity, sound and light discomfort EEG Rhythms (ed. Nunez, P. L.) 272–303 (Oxford of visual cortex excitability by five daily sessions thresholds and pre-attack state. Cephalalgia 20, University Press, 1995).
of repetitive transcranial magnetic stimulation 804–820 (2000).
NATURE REVIEWS NEUROLOGY VOLUME 10 MARCH 2014 153
2014 Macmillan Publishers Limited. All rights reserved 58. Lang, E., Kaltenhäuser, M., Neundörfer, B. of migraine. J. Headache Pain 6, 195–198 95. Llinás, R. R., Urbano, F. J., Leznik, E., & Seidler, S. Hyperexcitability of the Ramírez, R. R. & van Marle, H. J. Rhythmic and primary somatosensory cortex in migraine 76. De Marinis, M., Pujia, A., Natale, L., dysrhythmic thalamocortical dynamics: GABA —a magnetoencephalographic study. Brain 127, D'arcangelo, E. & Accornero, N. Decreased systems and the edge effect. Trends Neurosci. 2459–2469 (2004).
habituation of the R2 component of the blink 28, 325–333 (2005).
59. Oelkers-Ax, R., Parzer, P., Resch, F. & reflex in migraine patients. Clin. Neurophysiol. 96. Bjørk, M. H. & Sand, T. Quantitative EEG power Weisbrod, M. Maturation of early visual 114, 889–893 (2003).
and asymmetry increase 36 h before a migraine processing investigated by a pattern-reversal 77. de Tommaso, M. et al. Modulation of trigeminal attack. Cephalalgia 28, 960–968 (2008).
habituation paradigm is altered in migraine. reflex excitability in migraine: effects of attention 97. Bjørk, M. H. et al. The occipital alpha rhythm Cephalalgia 25, 280–289 (2005).
and habituation on the blink reflex. Int. J. related to the "migraine cycle" and headache 60. Sand, T., Zhitniy, N., White, L. R. & Stovner, L. J. Psychophysiol. 44, 239–249 (2002).
burden: a blinded, controlled longitudinal study. Visual evoked potential latency, amplitude and 78. Sand, T., Zhitniy, N., White, L. R. & Stovner, L. J. Clin. Neurophysiol. 120, 464–471 (2009).
habituation in migraine: a longitudinal study. Brainstem auditory-evoked potential habituation 98. Buchsbaum, M. & Silverman, J. Stimulus Clin. Neurophysiol. 119, 1020–1027 (2008).
and intensity-dependence related to serotonin intensity control and the cortical evoked 61. Sand, T., White, L., Hagen, K. & Stovner, L. Visual metabolism in migraine: a longitudinal study. response. Psychosom. Med. 30, 12–22 (1968).
evoked potential and spatial frequency in Clin. Neurophysiol. 119, 1190–1200 (2008).
99. Hegerl, U. & Juckel, G. Intensity dependence of migraine: a longitudinal study. Acta Neurol. 79. Siniatchkin, M., Kropp, P. & Gerber, W. D. auditory evoked potentials as an indicator of Scand. Suppl. 189, 33–37 (2009).
Contingent negative variation in subjects at risk central serotonergic neurotransmission: a new 62. Demarquay, G., Caclin, A., Brudon, F., Fischer, C. for migraine without aura. Pain 94, 159–167 hypothesis. Biol. Psychiatry 33, 173–187 & Morlet, D. Exacerbated attention orienting to auditory stimulation in migraine patients. 80. Paemeleire, K. & Schoenen, J. 31P-MRS in 100. Drake, M. E., Pakalnis, A. & Padamadan, H. Clin. Neurophysiol. 122, 1755–1763 (2011).
migraine: fallen through the cracks. Headache Long-latency auditory event related potentials 63. Omland, P. M. et al. Visual evoked potentials in 53, 676–678 (2013).
in migraine. Headache 29, 239–241 (1989).
interictal migraine: no confirmation of abnormal 81. Maniyar, F. H., Sprenger, T., Monteith, T., 101. Ambrosini, A., Coppola, G., Gérardy, P. Y., habituation. Headache 53, 1071–1086.
Schankin, C. & Goadsby, P. J. Brain activations in Pierelli, F. & Schoenen, J. Intensity dependence 64. Bednárˇ, M., Kubová, Z. & Kremlácˇek, J. Lack of the premonitory phase of nitroglycerin-triggered of auditory evoked potentials during light visual evoked potentials amplitude decrement migraine attacks. Brain 137, 232–241 (2014).
interference in migraine. Neurosci. Lett. 492, during prolonged reversal and motion stimulation 82. Chen, W. T. et al. Persistent ictal-like visual 80–83 (2011).
in migraineurs. Clin. Neurophys. cortical excitability in chronic migraine. Pain 152, 102. Proietti-Cecchini, A., Afra, J. & Schoenen, J. 254–258 (2011).
Intensity dependence of the cortical auditory 65. Coppola, G. et al. Lateral inhibition in visual 83. Chen, W. T. et al. Visual cortex excitability and evoked potentials as a surrogate marker of cortex of migraine patients between attacks. plasticity associated with remission from chronic central nervous system serotonin transmission J. Headache Pain 14, 20 (2013).
to episodic migraine. Cephalalgia 32, 537–543 in man: demonstration of a central effect for the 66. Sándor, P. S., Afra, J., Proietti-Cecchini, A., agonist zolmitriptan (311C90, Zomig). Albert, A. & Schoenen, J. Familial influences 84. Schoenen, J. Is chronic migraine a never-ending Cephalalgia 17, 849–854 (1997).
on cortical evoked potentials in migraine. migraine attack? Pain 152, 239–240 (2011).
103. Juckel, G., Hegerl, U., Molnár, M., Csépe, V. & Neuroreport 10, 1235–1238 (1999).
85. Currà, A. et al. Drug-induced changes in cortical Karmos, G. Auditory evoked potentials reflect 67. Lorenzo, C. D. et al. Cortical response to inhibition in medication overuse headache. serotonergic neuronal activity—a study in somatosensory stimulation in medication Cephalalgia 31, 1282–1290 (2011).
behaving cats administered drugs acting on overuse headache patients is influenced by 86. Di Clemente, L. et al. Interictal habituation deficit 5-HT autoreceptors in the dorsal raphe angiotensin converting enzyme (ACE) I/D genetic of the nociceptive blink reflex: an nucleus. Neuropsychopharmacology 21, polymorphism. Cephalalgia 32, 1189–1197 endophenotypic marker for presymptomatic 710–716 (1999).
migraine? Brain 130, 765–770 (2007).
104. Linka, T., Sartory, G., Gastpar, M., Scherbaum, N. 68. Restuccia, D., Vollono, C., Del Piero, I., 87. Coppola, G., Pierelli, F. & Schoenen, J. & Müller, B. W. Clinical symptoms of major Martucci, L. & Zanini, S. Somatosensory high Is the cerebral cortex hyperexcitable or depression are associated with the intensity frequency oscillations reflect clinical fluctuations hyperresponsive in migraine? Cephalalgia 27, dependence of auditory event-related potential in migraine. Clin. Neurophysiol. 123, 2050–2056 1427–1439 (2007).
components. Psychiatry Res. 169, 139–143 88. Knott, J. R. & Irwin, D. A. Anxiety, stress and 69. Restuccia, D., Vollono, C., Piero, I. D., the contingent negative variation. Arch. Gen. 105. Linka, T., Müller, B. W., Bender, S. & Sartory, G. Martucci, L. & Zanini, S. Different levels of Psychiatry 29, 538–541 (1973).
The intensity dependence of the auditory evoked cortical excitability reflect clinical fluctuations 89. Coppola, G. et al. Somatosensory evoked high- N1 component as a predictor of response to in migraine. Cephalalgia 33, 1035–1047 (2013).
frequency oscillations reflecting thalamo-cortical Citalopram treatment in patients with major 70. Kropp, P. & Gerber, W. D. Prediction of migraine activity are decreased in migraine patients depression. Neurosci. Lett. 367, 375–378 attacks using a slow cortical potential, the between attacks. Brain 128, 98–103 (2005).
contingent negative variation. Neurosc. Lett. 90. Sakuma, K., Takeshima, T., Ishizaki, K. & 106. Wang, W., Wang, Y. H., Fu, X. M., Sun, Z. M. & 257, 73–76 (1998).
Nakashima, K. Somatosensory evoked high- Schoenen, J. Auditory evoked potentials and 71. Siniatchkin, M., Kropp, P., Gerber, W. D. frequency oscillations in migraine patients. multiple personality measures in migraine and & Stephani, U. Migraine in childhood—are Clin. Neurophysiol. 115, 1857–1862 (2004).
post-traumatic headaches. Pain 79, 235–242 periodically occurring migraine attacks related 91. Lai, K. L., Liao, K. K., Fuh, J. L. & Wang, S. J. to dynamic changes of cortical information Subcortical hyperexcitability in migraineurs: 107. Sándor, P. S., Afra, J., Ambrosini, A. & processing? Neurosc. Lett. 279, 1–4 (2000).
a high-frequency oscillation study. Can. J. Neurol. Schoenen, J. Prophylactic treatment of migraine 72. Judit, A., Sándor, P. S. & Schoenen, J. Sci. 38, 309–316 (2011).
with beta-blockers and riboflavin: differential Habituation of visual and intensity dependence 92. Ambrosini, A., De Pasqua, V., Afra, J., Sándor, P. S. effects on the intensity dependence of auditory of auditory evoked cortical potentials tends to & Schoenen, J. Reduced gating of middle-latency evoked cortical potentials. Headache 40, 30–35 normalize just before and during the migraine auditory evoked potentials (P50) in migraine attack. Cephalalgia 20, 714–719 (2000).
patients: another indication of abnormal sensory 108. Burstein, R., Cutrer, M. F. & Yarnitsky, D. The 73. Evers, S., Quibeldey, F., Grotemeyer, K. H., processing? Neurosci. Lett. 22, 132–134 (2001).
development of cutaneous allodynia during Suhr, B. & Husstedt, I. W. Dynamic changes of 93. Coppola, G. et al. Interictal abnormalities of a migraine attack clinical evidence for the cognitive habituation and serotonin metabolism gamma band activity in visual evoked responses sequential recruitment of spinal and supraspinal during the migraine interval. Cephalalgia 19, in migraine: an indication of thalamocortical nociceptive neurons in migraine. Brain 123, 485–491 (1999).
dysrhythmia? Cephalalgia 27, 1360–1367 1703–1709 (2000).
74. Katsarava, Z., Giffin, N., Diener, H. C. & 109. Buchgreitz, L., Lyngberg, A. C., Bendtsen, L. & Kaube, H. Abnormal habituation of ‘nociceptive' 94. Llinás, R. R., Ribary, U., Jeanmonod, D., Jensen, R. Frequency of headache is related to blink reflex in migraine—evidence for increased Kronberg, E. & Mitra, P. P. Thalamocortical sensitization: a population study. Pain 123, excitability of trigeminal nociception. Cephalalgia dysrhythmia: a neurological and neuropsychiatric 19–27 (2006).
23, 814–819 (2003).
syndrome characterized by 110. Eide, P. K. Wind-up and the NMDA receptor 75. de Tommaso, M. et al. Habituation of single CO magnetoencephalography. Proc. Natl Acad. Sci. complex from a clinical perspective. Eur. J. Pain laser-evoked responses during interictal phase USA 96, 15222–15227 (1999).
4, 5–15 (2000).
154 MARCH 2014 VOLUME 10
2014 Macmillan Publishers Limited. All rights reserved 111. Mendell, L. M. & Wall, P. D. Responses of single 127. Valeriani, M. et al. Short-term plastic changes of 142. Stern, J., Jeanmonod, D. & Sarnthein, J. dorsal cord cells to peripheral cutaneous the human nociceptive system following acute Persistent EEG overactivation in the cortical pain unmyelinated fibres. Nature 206, 97–99 (1965).
pain induced by capsaicin. Clin. Neurophysiol. matrix of neurogenic pain patients. Neuroimage 112. Rhudy, J. L. et al. Pain catastrophizing is related 114, 1879–1890 (2003). 31, 721–731 (2006).
to temporal summation of pain but not temporal 128. Buchgreitz, L., Egsgaard, L. L., Jensen, R., 143. Stankewitz, A., Schulz, E. & May, A. Neuronal summation of the nociceptive flexion reflex. Pain Arendt-Nielsen, L. & Bendtsen, L. Abnormal brain correlates of impaired habituation in response 152, 794–801 (2011).
processing of pain in migraine without aura: to repeated trigemino-nociceptive but not to 113. Arendt-Nielsen, L., Brennum, J., Sindrup, S. & a high-density EEG brain mapping study. olfactory input in migraineurs: an fMRI study. Bak, P. Electrophysiological and psychophysical Cephalalgia 30, 191–199 (2010). Cephalalgia 33, 256–265 (2013).
quantification of temporal summation in the 129. Valeriani, M. et al. Nociceptive contribution to the 144. Mainero, C., Boshyan, J. & Hadjikhani, N. Altered human nociceptive system. Eur. J. Appl. Physiol. evoked potentials after painful intramuscular functional magnetic resonance imaging resting- Occup. Physiol. 68, 266–273 (1994). electrical stimulation. Neurosci. Res. 60, state connectivity in periaqueductal gray networks 114. Sandrini, G. et al. The lower limb flexion reflex in 170–175 (2008).
in migraine. Ann. Neurol. 70, 838–845 (2011).
humans. Prog. Neurobiol. 77, 353–395 (2005). 130. Valeriani, M. et al. Reduced habituation to 145. Russo, A. et al. Pain processing in patients with 115. You, H. J., Dahl, M. C., Chen, J. & experimental pain in migraine patients: a CO migraine: an event-related fMRI study during Arendt-Nielsen, L. Simultaneous recordings of laser evoked potential study. Pain 105, 57–64 trigeminal nociceptive stimulation. J. Neurol. wind-up of paired spinal dorsal horn nociceptive 259, 1903–1912 (2012).
neuron and nociceptive flexion reflex in rats. 131. Lev, R., Granovsky, Y. & Yarnitsky, D. Enhanced 146. Tessitore, A. et al. Interictal cortical Brain Res. 960, 235–245 (2003).
pain expectation in migraine: EEG-based reorganization in episodic migraine without aura: 116. Arendt-Nielsen, L. et al. The effect of N-methyl-d- evidence for impaired prefrontal function. an event-related fMRI study during parametric aspartate antagonist (ketamine) on single and Headache 53, 1054–1070 (2013).
trigeminal nociceptive stimulation. Neurol. Sci. repeated nociceptive stimuli: a placebo 132. de Tommaso, M. et al. Lack of habituation 32 (Suppl. 1), S165–S167 (2011). controlled experimental human study. of nociceptive evoked responses and pain 147. Szabó, N. et al. White matter microstructural Anesth. Analg. 81, 63–68 (1995).
sensitivity during migraine attack. alterations in migraine: a diffusion-weighted MRI 117. Serrao, M. et al. Effects of diffuse noxious Clin. Neurophysiol. 116, 1254–1264 (2005).
study. Pain 153, 651–656 (2012).
inhibitory controls on temporal summation of 133. de Tommaso, M., Marinazzo, D. & Stramaglia, S. 148. Burstein, R. et al. Thalamic sensitization the NWR reflex in humans. Pain 112, 353–360 The measure of randomness by leave-one-out transforms localized pain into widespread prediction error in the analysis of EEG after laser allodynia. Ann. Neurol. 68, 81–91 (2010). 118. Perrotta, A. et al. Sensitisation of spinal cord painful stimulation in healthy subjects and 149. Coppola, G. et al. Dynamic changes in thalamic pain processing in medication overuse migraine patients. Clin. Neurophysiol. 116, microstructure of migraine without aura patients: headache involves supraspinal pain control. 2775–2782 (2005).
a diffusion tensor magnetic resonance imaging Cephalalgia 30, 272–284 (2010).
134. Ayzenberg, I. et al. Central sensitization of the study. Eur. J. Neurol. 119. Perrotta, A. et al. Oral nitric-oxide donor glyceryl- trigeminal and somatic nociceptive systems in trinitrate induces sensitization in spinal cord medication overuse headache mainly involves 150. Viganò, A. et al. Transcranial Direct Current pain processing in migraineurs: a double-blind, cerebral supraspinal structures. Cephalalgia 26, Stimulation (tDCS) of the visual cortex: placebo-controlled, cross-over study. Eur. J. Pain 1106–1114 (2006).
a proof-of-concept study based on interictal 15, 482–490 (2011).
135. Antal, A., Arlt, S., Nitsche, M. A., Chadaide, Z. electrophysiological abnormalities in migraine. 120. Yarnitsky, D. et al. Recommendations on & Paulus, W. Higher variability of phosphene J. Headache Pain 14, 23 (2013).
terminology and practice of psychophysical DNIC thresholds in migraineurs than in controls: 151. Teepker, M. et al. Low-frequency rTMS of the testing. Eur. J. Pain 14, 339 (2010).
a consecutive transcranial magnetic stimulation vertex in the prophylactic treatment of migraine. 121. Perrotta, A. et al. Acute reduction of anandamide- study. Cephalalgia 26, 865–870 (2006).
Cephalalgia 30, 137–144 (2010).
hydrolase (FAAH) activity is coupled with a 136. Antal, A. et al. Homeostatic metaplasticity of the 152. Antal, A. et al. Cathodal transcranial direct reduction of nociceptive pathways facilitation motor cortex is altered during headache-free current stimulation of the visual cortex in the in medication-overuse headache subjects after intervals in migraine with aura. Cereb. Cortex 18, prophylactic treatment of migraine. Cephalalgia withdrawal treatment. Headache 52, 2701–2705 (2008).
31, 820–828 (2011).
1350–1361 (2012).
137. Siniatchkin, M. et al. Abnormal changes of 153. Marinazzo, D., Liao, W., Chen, H. & 122. Giffin, N. J., Katsarava, Z., Pfundstein, A., synaptic excitability in migraine with aura. Stramaglia, S. Nonlinear connectivity by Granger Ellrich, J. & Kaube, H. The effect of multiple Cereb. Cortex 22, 2207–2216 (2012).
causality. Neuroimage 15, 330–338 (2011).
stimuli on the modulation of the ‘nociceptive' 138. Brighina, F. et al. Abnormal facilitatory blink reflex. Pain 108, 124–128 (2004).
mechanisms in motor cortex of migraine with 123. Bromm, B. & Treede, R. D. Nerve fibre aura. Eur. J. Pain 15, 928–935 (2011).
M.d.T., A.A., F.B., G.C., A.P., F.P., G.S. and M.V. discharges, cerebral potentials and sensations 139. Pierelli, F., Iacovelli, E., Bracaglia, M., Serrao, M. participated in writing this Review on behalf induced by CO laser stimulation. & Coppola, G. Abnormal sensorimotor plasticity of the Italian Group for Neurophysiology of Migraine, Hum. Neurobiol. 3, 33–40 (1984).
in migraine without aura patients. Pain 154, created by members of the Italian Society for the 124. Garcia-Larrea, L., Frot, M. & Valeriani, M. Brain 1738–1742 (2013).
Study of Headaches and the Italian Society of generators of laser-evoked potentials: from 140. Adjamian, P., Sereda, M., Zobay, O., Hall, D. A. dipoles to functional significance. & Palmer, A. R. Neuromagnetic indicators of Neurophysiol. Clin. 33, 279–292 (2003).
tinnitus and tinnitus masking in patients with Author contributions 125. de Tommaso, M. et al. Topographic and dipolar and without hearing loss. J. Assoc. Res. All authors researched data for and participated in analysis of laser-evoked potentials during Otolaryngol. 13, 715–731 (2012).
writing of the article. In addition,M.d.T., A.A., G.C., F.P. migraine attack. Headache 44, 947–960 (2004).
141. Walton, K. D., Dubois, M. & Llinás, R. R. and J.S. contributed to discussion of content, and 126. de Tommaso, M. et al. Changes in cortical Abnormal thalamocortical activity in patients M.d.T., A.A. and J.S. contributed to reviewing and/or processing of pain in chronic migraine. with complex regional pain syndrome (CRPS) editing of the manuscript before submission. M.d.T. Headache 45, 1208–1218 (2005). type. Pain 150, 41–51 (2010).
and A.A. contributed equally to this manuscript.
NATURE REVIEWS NEUROLOGY VOLUME 10 MARCH 2014 155
2014 Macmillan Publishers Limited. All rights reserved

Source: http://www.amicacard.it/data/convenzionanti/upload/63068/studio%20Nature_emicrania.pdf