Visual cortex excitability and plasticity associated with remission from chronic to episodic migraine

Visual cortex excitability varies with clinical migraine status

The main finding of this study is the significant change in MEG measurement of visual cortex excitability when CM remits to EM. The P100m parameters featuring CM, including increased amplitude at the initial block and response habituation to repetitive stimulations, both shifted towards those characterizing EM, i.e. decreased initial amplitude with subsequent potentiation. The significant modification indicates that visual cortex excitability is not a steady physiological trait inherent in CM; rather, it may switch in tandem with the clinical status of migraine (CM vs EM).

The modifiable excitability in CM remission is in line with a TMS study which used magnetic suppression of perceptual accuracy (MSPA) to show altered visual cortex excitability when headache frequency was significantly reduced in CM by topiramate; however, the finding was not independent of a direct pharmacological effect of topiramate due to the lack of a washout period (19). It is also notable that CM, despite potential intractability, still remains a relatively fluid clinical status of migraine in epidemiological studies (12,13). In contrast with the one-third of CM patients maintaining the same diagnosis throughout, one-fourth experienced CM remission at the biannual follow-ups in the American Migraine Prevalence and Prevention study (13).

Changes of visual cortex excitability in relation to CM pathophysiology

The parallel switch of excitability in CM remission further consolidates the pathophysiological link of CM to visual cortex excitability. The reduction of P100m amplitudes after CM remission, especially at the initial block, is compatible with the resolution of visual cortex sensitization that underpins exaggerated P100m amplitudes in CM as shown earlier (9). On the other hand, the finding of a lack of habituation here, when regarded as up-regulated excitability, seems contradictory to the above amplitude decrease which suggests ‘normalized’ hyperexcitability. Nevertheless, this may reflect a resolution of continuous ictal activation in CM due to decreased central sensitization, and a resumed central excitability that features EM, hence an interictal lack of normal habituation. Taken together, the present findings are physiologically in agreement with the shattered steady-state ictal activation and regaining of the dynamic modulation of central excitability after CM remission.

Of note, the reduction of P100m amplitude after CM remission did not approximate the low level as shown in the interictal EM patients in our earlier study (35.6 ± 2.9 nAm in grand average, 33.5 ± 3.0 nAm at block 1) (9). The degree of potentiation (sub-averaged amplitude changes) in remitted CM was also smaller than in the EM patients in our earlier study (Figure 2). The exact reason for such a lesser extent of excitability change in remitted CM than in EM is unknown. Despite the possibility of topiramate effect, as discussed in the next section, a straightforward explanation is that remission was still in its early stage and these ‘remitted’ patients still had high frequency EM (monthly headache days 8.7 ± 2.3 vs 6.5 ± 5.2 in the EM patients in our earlier study) (9).

Studies using positron emission tomography (PET) have revealed ictal-like neural dysfunction in interictal CM because the brainstem activation shown in these studies was also demonstrated during acute migraine (20–23). It was intriguing that brainstem activation persisted when head pain associated with CM was completely suppressed by the short-term use of suboccipital neuromodulation (22). Therefore, ictal-like central activation is independent of headache severity and may linger during transient (or early) remission of CM. The difference in excitability measures between remitted CM (here) and EM (9) might be explained by lingering ictal activity, given the featured excitability change during ictal periods. Nevertheless, the inference needs confirmation by multimodal neuroimaging (or neurophysiological) studies after long-term remission and a significant improvement in headache frequency, and pertinent findings may further explain why the ictal-like MEG activity, as shown in our earlier CM study (9), is independent of the presence or absence or the intensity of headache at the time of recording (11).

The effect of topiramate

Another possibility for the present excitability change is the effect of topiramate. Indeed, central excitability in migraine may be altered by topiramate (24,25). However, the pharmacological effect, or conversely, the withdrawal effect of topiramate as the culprit of the complex excitability changes reported here is not likely. The reasons are threefold. First, the excitability-modulating effect of topiramate seems to be dose-dependent (26,27). A study in patients with chronic neuropathic pain showed that facilitation of muscle-evoked potentials (MEPs) by repetitive TMS can be blocked by topiramate treatment at 75 and 100 mg/day but not at lower doses (25 and 50 mg/day) (27). Of note, the use of topiramate in the present study was tapered then discontinued at least 3 days before MEG follow-up. The majority (nine out of 10) of the CM patients had a washout period at least 5 days before MEG follow-up (except for a 3-day washout in patient G), and the overall washout period averaged 8.9 ± 6.1 days, in contrast with the estimated 19–23 h half-life of topiramate (28). Taken together, the pharmacological effect of topiramate at MEG follow-up, if any, might be negligible in terms of pharmacokinetics, although the threshold level at which topiramate alters visual cortex excitability is undetermined.

Second, the present findings of complex excitability changes seem inexplicable by the inhibitory effect of topiramate. To date, the exact mechanism of topiramate in preventing migraine remains elusive; however, studies on excitability modulation (6) showed that topiramate down-regulates cerebral excitability via reduced central excitation or enhanced central inhibition (19,24–27,29). Although reduced P100m amplitude in CM remission is compatible with down-regulated excitability, the finding of increased sub-averaged amplitude changes seems contradictory. Given the earlier finding that topiramate reversed dishabituation of contingent negative variation in migraine (24), the sub-averaged amplitude changes should be decreased by topiramate. Nevertheless, it is intriguing to note that the inhibitory effect of topiramate, if present after washout, may be an alternative explanation for a lesser extent of excitability change in remitted CM compared to our earlier findings in EM (Figure 2).

Finally, in support of the above assumptions against topiramate effect, our unpublished data did not show any difference in P100m parameters between subjects with 7 (n = 6) vs ≥ 7 (n = 4) days of topiramate washout; furthermore, two subjects (patients A and H; see Figure 1) who were followed after discontinuing topiramate for around 2 weeks, underwent a second follow-up MEG after a 1-month washout and showed a similar increase in the sub-averaged amplitude change at block 30.

Lack of a clinical predictor of excitability changes

It is not unexpected that this study failed to find any clinical predictor of the longitudinal excitability change, given the lack of any cross-sectional correlation between MEG results and clinical parameters (9). A TMS study using the MSPA technique also did not yield a clinical/physiological correlation in CM patients treated with topiramate (19). Nevertheless, reduction of headache frequency correlated inversely with an increase of phosphene thresholds in another TMS study treating EM patients with topiramate (29). Besides different methodologies, the complex pathophysiology of CM may thus be the culprit here for the lack of a clinical/physiological correlation.

Limitations of the study

This study was constrained by a small sample size because the MEG follow-up was dependent on topiramate responsiveness. Since this study focused on the changes in visual cortex excitability associated with CM remission, rather than the different excitability patterns between treatment responders and non-responders, we did not follow up MEG in those CM patients who failed to remit to EM. Of note, these ‘intractable’ CM patients might provide further information deciphering potential impacts on MEG measurement resulting from topiramate or migraine pain. Without the comparison with these treatment non-responders, the present finding based on treatment responders might not confirm visual cortex excitability as a faithful biomarker of CM evolution/remission.

However, there were ethical concerns of headache rebound and treatment delay with the discontinuation and washout of topiramate in these patients, for the sake of an interictal MEG follow-up. Moreover, generalization of the present data is justified by comparable demographics, headache profiles and MEG measurements between the CM patients with (n = 10) and without (n = 15) follow-up (Table 1), and a lack of statistical difference in baseline MEG between topiramate responders (n = 10) and non-responders (n = 7) (data not shown). Further studies are needed to clarify if the present findings can be generalized to those CM patients with long-term or spontaneous remission, or those who undergo different preventive treatments, such as onabotulinumtoxinA. It is uncertain whether the excitability change found here is confined to the visual modality or is pervasive across different sensory modalities. The limitation of hemifield visual stimulation has been discussed earlier (6,9).