Electrophysiological correlates of episodic migraine chronification: evidence for thalamic involvement

Migraine is one of the most prevalent and disabling neurological disorders [1, 2]. It is characterized by recurrent attacks of headache widely variable in duration and intensity, usually accompanied by nausea/vomiting and/or photo-/phono- phobia. In some cases, migraine patients experience a progressive increase in the frequency of the attacks, leading to headache chronification. This clinical condition is defined as 15 or more headache days with 8 or more migraine attacks per month [3]. The exact pathophysiological mechanisms underlying chronic migraine are still under intense scrutiny. Possible culprits for pain chronification include central sensitization and defective central pain control systems [4, 5].

During the last decades, clinical neurophysiology methods have allowed in vivo measurements of the migraineur’s electrocortical responses to various sensory stimuli. Altered thalamo-cortical connections [6, 7], with cortical dysexcitability [8] and lack of habituation in response to various sensory stimuli, characterize episodic migraineurs’ brains [9]. This abnormal information processing increases during the pain-free days, reaching its maximum just before the attack onset, and disappears in the ictal phase [6, 10‐16]. Less is known about how mechanisms underlying headache chronification alter this electro-functional profile in episodic patients experiencing a conversion to CM [17].

Evidence has been recently found in favour of persistent somatosensory system central sensitization in chronic headache secondary to medication overuse (medication overuse headache, MOH) by recording cortical somatosensory evoked potentials (SSEPs) [10, 18]. To the best of our knowledge, no study has investigated simultaneously SSEP habituation as well as thalamo-cortical connections in episodic migraine patients who developed chronic migraine without the confounding factor of medication overuse. Having this information may contribute to shed more light on the mechanisms underlying headache transformation.

With this specific purpose, we designed the present study to explore whether the sensory cortical response pattern differs between CM patients and patients with episodic migraine without aura recorded both during and between attacks. To do so we recorded low-frequency (LF) SSEP in order to assess habituation/sensitization phenomena. Thereafter, we studied high-frequency oscillations (HFOs) embedded in the common SSEPs in order to identify two bursts of HFOs: an early component thought to be generated by pre-synaptic thalamo-cortical afferents, and a late component reflecting post-synaptic cortical activation [19‐21]. We also sought possible correlations between the electrophysiological pattern and clinical features including the number of days with headache and the duration of the chronification phase.

Subjects - Among consecutive patients attending our headache clinic, 53 patients gave informed consent to participate in the study (Table 1) which was approved by the local ethics committee. According to the new ICHD-III criteria [3], 19 patients were diagnosed as having chronic migraine (CM) during their first visit. These patients were not affected by medication overuse since the mean monthly tablet intake was 2.7 ± 3.0. Before progressing to CM, all patients had a clear-cut history of episodic migraine without aura (MO, ICHD-II code 1.1). With the exception of 2 patients who had a mild headache (mean VAS = 3), all CM patients underwent the SSEP recordings in a pain-free state. The 2 patients who had a headache had no associated migrainous features. Thirty-four patients were diagnosed as having episodic migraine without aura (ICHD-II code 1.1). Of these, 15 were recorded during the interictal period (MOii), i.e. at least three days before and after an attack, and 19 within a time range of 12 hours before or after the beginning of an attack, then considered as ictal (MOi). Neither chronic nor episodic patients were allowed to take any prophylactic medication in the three months before the recording session. For those patients experiencing a migraine attack, no acute anti-migraine drugs were allowed until the end of the recording session. For comparison, we enrolled 20 healthy volunteers (HV) of comparable age and sex distribution; they had no personal or familial history (1st or 2nd degree relatives) of migraine or any detectable medical conditions. To avoid variability due to hormonal changes, women were recorded outside their pre-menstrual or menstrual periods. All participants received a complete description of the study and granted written informed consent. The project was approved by the ethical review board of the Faculty of Medicine, University of Rome, Italy.

Assessable SSEP recordings were obtained from all patients and controls participating in the study (explicative traces in Figure 1).

We designed this study to investigate how headache chronification alters subcortico/cortical somatosensory response patterns of episodic migraine patients experiencing progression to chronic migraine without medication overuse.

In our episodic migraineurs between attacks, we have confirmed previous results showing, on the one hand, that the LF-SSEP N20-P25 amplitude lacks habituation during stimulus repetition despite an initially low response amplitude [10, 22], and that this comes in parallel to a reduced somatosensory thalamocortical activity [6, 23], as reflected by the low amplitude of the early HFO, on the other hand.

Chronic migraine patients have shown a neurophysiological pattern quite similar to that of episodic migraineurs recorded during an attack. In fact, both groups of patients were characterized by higher amplitudes after low numbers of median nerve electrical stimuli (block 1), reflecting sensory cortex sensitization, and by response habituation over sequential block averages. This combination of subcortico/cortical electrophysiological patterns observed in our chronic migraineurs was previously defined as a condition of “never-ending migraine attack” [22]. Moreover, after band-pass filtering, in both CM and MOi patients the interictal episodic amplitude reduction normalized, besides the amplitudes of the primary cortical component consistently increasing with respect to HV and MOii.

In migraine, a strong relationship between clinical and neurophysiological profiles has been demonstrated previously. In episodic migraine, several studies have confirmed between attacks that the amplitude of sensory evoked potentials decreased or tended to be decreased for low numbers of delivered stimuli and lack of habituation during subsequent block averages [9, 10, 22]. Contrariwise, close to or during an attack, initial SSEP amplitude increased and deficit of habituation normalized [10, 13, 16] demonstrating that, on the one hand, the cortical activation state changes between the interictal and ictal periods, and that sensitization disappears between attacks, on the other hand. The electrocortical pattern we found here in our CM patients may thus suggest that the sensory cortex is persistently locked in an “ictal”-like state, associating with both sensitization and habituation. This pattern contrasts with that of episodic migraine where the transition between two electrocortical states (habituation and its lack) alternates following the recurrence of the migraine cycle (ictal and interictal). Similar results were recently observed in a visual EP study [11]. Moreover, this electrophysiological pattern partially differs from that found in chronic headache due to medication overuse (MOH) where, after a similar initial increase in response amplitude, a further increase was observed during stimuli repetitions [10]. Therefore, we considered MOH patients as locked in a “pre-ictal” state, in analogy with the cortical response pattern detected a few days before the beginning of an attack in episodic migraine [15, 16]. Nevertheless, further underscoring the clinico-electrophysiological inter-relationship in migraine, this phenomenon in MOH evolved from migraine, was strongly dependent on the drug of overuse, since it was maximal in patients overusing NSAIDs and almost non-existent in those who overuse only triptans [10].

From a behavioural point of view, two distinct and independent processes govern the outcome of repetitive stereotyped sensory stimulation: an increasing one called sensitization, and a reducing one called habituation [24, 25]. Sensitization occurs at the beginning of the test session and is responsible for the transitory increase in response amplitude, whereas habituation occurs throughout the test session, and is responsible for the late response decrement [26]. From a physiological point of view, demonstrative of central sensitization are the plastic changes in neural networks devoted to the processing of pain information [27] that results in abnormal neural excitability with decreased nociceptive thresholds and increased responsiveness to noxious and usually innocuous peripheral stimuli, such as, in our case, somatosensory ones [28]. Experimental studies in animals [29] and humans [30] show that SSEP amplitudes increase when transient intense activation of nociceptive afferents induces central sensitization, as happens in clinical pain conditions including chronic headache. Our study shows that sensitization, as reflected by increased initial SSEP amplitudes, is equally present in CM and MOi, although we managed not to record CM patients during a full-blown migraine attack. In CM and episodic migraine attacks central sensitization is associated with increased excitability both at the trigeminal system [31, 32] and the supraspinal levels [33, 34]. Interestingly, from our correlation analysis it emerges that the more pronounced the habituation, the more the CM number of headache days and sensory sensitization increase. This further reinforces the evidence for a strong relationship between clinical and neurophysiological features in migraine. A well-recognized clinical expression of central sensitization is cutaneous allodynia, which was shown to be prevalent during episodic migraine attacks at cephalic and extracephalic sites [35, 36]. This phenomenon is even more evident in chronic migraine [37‐39]. Based on animal models of experimental pain [40, 41], some hypothesized that temporary sensitization of third-order thalamic neurons receiving convergent input from the dura, periorbital skin, and skin areas at different body sites explains the spread of cutaneous allodynia beyond the initial pain area during an attack of migraine [36, 42]. In a recent study, Burstein and colleagues (2010) studied the effects of sensitizing skin stimuli on the activity of third-order trigeminovascular neurons in the rat thalamus, and on thalamic activation registered by fMRI during migraine in humans. On the one hand, the rat thalamus exhibited long-lasting hyperexcitability to cephalic and extracephalic skin stimuli in response to sensitizing inflammatory soup, and patients undergoing migraine experienced acute thalamic activation during the fMRI in response to extracephalic brush and heat stimuli, on the other hand [42]. The latter data indicate that allodynia is associated with sensitization of third-order thalamic neurons. Taking into account the latter evidence, our finding that the migraine attack modifies a neuronal activity (early HFO) that is generated in the thalamus is not surprising. We observed in both CM and episodic migraineurs recorded ictally an amplitude increase of the usually low thalamo-cortical activation (early HFO) of the interictal episodic migraineurs in parallel with a rise in primary cortical activation (late HFO). These findings may contribute to shed light on the mechanisms of central sensitization, since the elusive mechanisms that are able to ignite the cascade of events that finally lead to an attack facilitate the thalamic activity by reducing latency and augmenting amplitude and, in turn, enhance primary cortical activation at the low (1st SSEP amplitude block) and high frequency bands (late HFOs). This interpretation is supported by the direct correlation we found between the amplitude of the early and that of the late HFO components: in other words, the more the amplitude of the thalamic component increased, the more the cortical activation was enhanced, especially in CM and MOi patients. Furthermore, from the Pearson’s analysis it also emerged that the more this sequential thalamic and cortical HF oscillatory activation increased, the more marked the sensitization of the LF-SSEP N20-P25 1st amplitude block.

Only hypotheses can be made about the neural mechanisms by which thalamic neurons become facilitated in CM and in MOi. This process may primarily involve sequential sensitization of first-order or second-order trigeminovascular neurons probably through an evident (aura) or silent (without aura) cortical spreading depression or, more likely, through an indirect activation of pain modulatory structures in the brainstem (raphe magnus, locus coeruleus and other aminergic nuclei) and the forebrain (periaqueductal gray, rostroventral medulla) [43, 44]. That the brainstem plays a relevant role in mediating and maintaining central sensitization was recently confirmed by functional neuroimaging studies in healthy humans [45, 46]. In Migraine, clear examples of monoaminergic brainstem activation come from both neurophysiology [12, 47] and neuroimaging studies where the dorsal rostral brainstem, which contains state-setting aminergic, including serotonergic, nuclei projecting to the thalamus and cortex, was activated immediately before [48] and during an attack of episodic [49, 50] and chronic [4, 51] migraine.

In conclusion, we show for the first time that not only is the response pattern of the somatosensory cortex in CM patients similar to that found during a migraine attack in episodic migraine patients, in such a way that both habituate normally, as previously observed with visual responses [52], but also that they show an initial response sensitization. Moreover, from the analysis of high-frequency oscillations, it clearly emerges that this sensory sensitization may be explained by the fact that in both groups, the connections between the thalamus and cortex intensify compared to episodic migraine between attacks. Whether this electro-functional behaviour is primary or secondary to daily headache, thus reflecting an electrophysiological fingerprint of the somatosensory system central sensitization process remains to be determined.