Abstract
Background: Results from transcranial magnetic stimulation (TMS) studies of visual cortex have confirmed visual cortical hyperexcitability in patients with migraine. It has been speculated that this may be due to deficient intracortical inhibitory tone. However, the TMS induction of phosphenes relies on the reporting of a subjective experience, and may thus be subject to bias.
Methods: Seven migraineurs with visual aura and seven sex- and age-matched controls were studied. Fifty-four different three-letter combinations were briefly displayed and followed by a magnetic pulse at 40, 70, 100, 130, 160, and 190 msec. Subjects were required to report as many letters as they thought they had recognized.
Results: In the migraine group, the mean proportion of correctly identified letters was significantly higher at 100 msec, as was the proportion of trials with two or three letters correctly reported. The time window in which perceptual suppression could be introduced was narrower in migraineurs compared to controls.
Conclusion: These findings suggest that inhibitory systems are activated to a lesser extent by TMS pulses in patients. This observation is in agreement with the hypothesized deficiency of intracortical inhibition of the visual cortex, at least in migraineurs with aura.
Migraine is a paroxysmal headache disorder, the cause of which is unknown. It has been proposed that trigeminovascular as well as central serotonergic, noradrenergic,1 and possibly even dopaminergic2 modulatory systems may be involved in the pathogenesis of the head pain. Moreover, researchers are turning to the idea that the cerebral cortex may play a pivotal role early on in the sequence of events leading to a migraine attack, even in migraine without aura.3 Several studies have shown that the cerebral cortex in patients with migraine displays an enhanced responsiveness to various external stimuli.4-11⇓⇓⇓⇓⇓⇓⇓ This cortical neuronal hyperexcitability has been postulated as an important contributing factor in the pathogenesis of migraine.12
One possible mechanism for this supposed cortical hyperexcitability may be a lack of inhibitory control in the primary visual cortex (V1). Previous research into functional changes has demonstrated differences between patients with migraine and control subjects on certain visual tasks, which are believed to depend on GABA-mediated inhibition in visual cortex.13 For instance, it has been reported that patients with migraine with aura admit more illusions and greater visual discomfort than headache free controls when viewing high contrast square wave gratings (striped patterns).14 Recent work in our laboratory using the method of metacontrast masking15 has shown that perceptual suppression of a simple target (presented on a computer screen) by a subsequently presented and spatially nonoverlapping mask is less effective in migraine with aura.16 In primates, this type of perceptual suppression is known to be dependent upon inhibitory interactions at the level of the primary visual cortex17; thus, our observation that metacontrast masking is lessened in migraine with aura is highly consistent with a decrease in the functioning of inhibitory neuronal systems in the primary visual cortex of these patients.
Transcranial magnetic stimulation (TMS) of the visual cortex has been one of the tools used to demonstrate increased cortical excitability in migraineurs, showing that thresholds for the induction of magnetophosphenes are consistently reduced in migraineurs.8-11⇓⇓⇓ Although the magnetophosphene induction method is an important first step toward understanding cortical dysfunction in migraine, it is somewhat prone to artifact, as we have mentioned elsewhere.18 In this study, we sought to explore other paradigms that permit a more detailed examination of neuronal processes, utilizing objective dependent measures. It has been demonstrated that veridical perception of a visual stimulus can be suppressed by a single TMS pulse administered 60 to 120 msec after the stimulus presentation,19 in a fashion rather similar to the metacontrast masking technique mentioned above. It has been proposed that this effect is the net result of an enhancement of inhibitory mechanisms, either by direct action of the magnetic pulse on cortical neurons20 or by the induction of inhibitory postsynaptic potentials by the pulse.21-23⇓⇓ In an attempt to evaluate further the state of the occipital inhibitory neuronal systems in migraineurs, we therefore applied the TMS-suppression-of-perception design in a sample of migraineurs with aura and an age- and sex-matched control group. By direct analogy with the results from our metacontrast masking study, we predicted that suppression of visual perception by a TMS pulse would be lessened in migraineurs as compared to controls.
Methods.
Subjects.
Seven patients with migraine with visual aura (MA) (six female; one male) and seven headache-free age- and sex-matched controls (C) participated in the study. Subjects were recruited from the Headache Outpatient Clinic of the Neurology Department at the Atrium Medical Center, Heerlen, the Netherlands, and from Lancaster University, Lancaster, UK. All patients were seen by a neurologist experienced in headache diagnosis (W.M.M.) and were classified according to the criteria of the International Headache Society.24 All patients with migraine reported that they experienced visual aura on at least 50% of their migraine attacks. None of the control subjects had any family history of migraine. The mean age of the MA patients was 34.43 years (SD: 12.04; range: 18 to 55) and the C subjects 35.71 years (SD: 12.62; range: 23 to 60).
Subjects were not eligible for study participation if magnetic stimulation was considered unsafe, i.e., in case of epilepsy or attachment of electronic or metal objects to or in the body. Moreover, neurologic or ophthalmologic conditions other than refractive error were not allowed, nor was the use of prohibited concomitant medication (antidepressants, minor and major tranquilizers, lithium, anticonvulsants, antiparkinsonians, muscle relaxants, systemic anticholinergics, migraine prophylactics, Ca-entry blockers, antiemetics, betahistine, cinnarizine, piracetam, hormone replacement therapy) in the month prior to the study. Acute migraine medication, not containing opioids, was permitted but not in the 24 hours prior to the assessment. Ethical approval for the study was obtained from the ethical committees of Atrium Medical Center, Heerlen, the Netherlands, and from the Department of Psychology, Lancaster University, Lancaster, UK. Written informed consent was obtained from all participants before testing.
Stimuli and apparatus.
All subjects had normal (corrected) visual acuity as assessed by Snellen chart.
Visual target stimuli consisted of low contrast letter trigrams presented centrally within a frame ( figure 1). The letters were presented in upper case Helvetica, font size 48. The presence of a frame helped to equalize any crowding effect on the letters and therefore also improve their legibility.25 The letters used were chosen from a subset of letters of approximately equal legibility.26 The letter trigrams, within the frame, subtended 1.73° × 0.79° of visual angle when viewed from a distance of 175 cm. The mean luminance of the stimuli was 24 cd/m2 whereas the mean luminance of the background was 30 cd/m2. There was no additional color contrast between the trigrams and the background: they differed only on a calibrated gray scale. Visual stimuli were presented on an Apple Macintosh monitor driven by a Macintosh Power Mac computer, running SuperLab (Cedrus Corp., Phoenix, AZ) software.
Figure 1. Example of displayed letter trigram.
TMS was conducted using a MagStim 200 (The MagStim Company, Cardiff, UK). This has a maximal output of 2 T when used with a 90-mm circular coil. Subjects wore a tightly fitting EEG cap, marked with an orientation grid at the back and stripped of its metal-containing electrodes and wiring. The grid was placed symmetrically over the occipital area of the scalp. Inion, nasion, and preauricular points were used as reference points for the appropriate fitting of the cap. The orientation grid covered a rectangular area ranging from 4 to 10 cm superior to the inion and 4 cm from the midline on either side; stimulation coordinates were marked on the EEG cap at 2-cm intervals along X- and Y-axis.
Procedure.
All patients were tested interictally at least 24 hours after the last migraine attack. The letter trigrams were presented for 40 msec and subjects were asked to report verbally the letters in the correct order, which were then recorded by an experimenter (subjects were asked to say “blank” or “don’t know” if they were unaware of a letter at a specific position). A magnetic pulse was given following presentation of the letters at variable intervals. Testing consisted of two phases.
Phase one—threshold determination.
Before magnetic stimulation began subjects completed a series of practice trials. The first three trigrams were presented for 250 msec to familiarize subjects with the stimuli. Subjects then completed 10 practice trials at 40 msec. If subjects were unable to report the letters accurately by the end of this session, the practice trials were rerun to ensure that subjects were familiar with the procedures and were able to perceive the stimuli accurately. Following this, one of the patients with migraine was still unable to report the letters accurately. Consequently, the contrast of the letters was increased slightly (background luminance was raised to 37 cd/m2) for this subject only so that her performance at baseline was equal to that of all other subjects.
Next, a TMS pulse was triggered at a fixed interval of 100 msec after the onset of the target trigram (stimulus onset asynchrony [SOA]), observing an interval of at least 5 seconds between successive magnetic pulses. Previous studies19,27⇓ and our own pilot experiments have demonstrated that the 100 msec interval provides peak suppression in subjects. Magnetic stimulation started at 50% stimulus intensity and was increased in 10% steps until the subject was unable to identify at least two of the three target letters correctly in the order presented (≤33% accuracy). Once this was achieved, the level of stimulation was fine-tuned. Then, without changing the stimulator output, the coil was moved around the grid to exclude a coordinate with a lower intensity at criterion. In our own studies of phosphene thresholds it has been shown that the interindividual variability of the lowest-threshold coordinate requires that the occipital area must be scanned to determine the optimal stimulation point.11 This is consistent with the high interindividual variability of the amount and distribution of striate cortex on the surface of the occipital pole.28 As phosphenes and suppression of visual perception may be produced in the same part of the visual cortex,29 it was considered prudent to apply similar methodologic standards.
Phase two—time course of suppression.
In phase two, subjects were presented with 54 trials in which the letter trigrams were followed at a variable interval by the magnetic pulse. Six SOA were tested—40, 70, 100, 130, 160, and 190 msec—with subjects completing nine trials at each SOA. The order of presentation of trials was randomized for each subject. The intensity of the TMS pulse and point of stimulation was set as determined in phase one. Throughout, an interval of at least 5 seconds between successive magnetic pulses was observed. Subjects responded verbally by trying to name the letters in the order as they were presented. An experimenter recorded their responses.
Results.
The mean output of the stimulator (expressed as percentage of maximum output) required for successful suppression was not significantly different between the groups (MA: mean 78.43, SD 8.79; C: mean 76.43, SD 6.95; t(12) = 0.472; p = 0.65). For each individual, the proportion of letters correctly identified (PC) at each SOA (i.e., number correct as a proportion of the total of 27 letters presented over nine trials at each of the six intervals) was calculated. Group mean PCs, obtained during the time course phase, were then compared using a diagnosis (MA; C) by SOA (40, 70, 100, 130, 160, 190 msec) mixed analysis of variance (ANOVA), and diagnosis being the only between group factor ( table, figure 2). There was a main effect of diagnosis, in that the MA patients identified more letters correctly than the C subjects (F(1,12) = 7.03, p < 0.05). There was also a main effect of SOA: suppression of perceptual accuracy was least at SOA 40, 160, and 190 msec, whereas suppression peaked at 100 msec SOA (F(5,60) = 11.08, p < 0.01). This is consistent with previous studies19,27⇓ that have shown that suppression is maximal at approximately 100 msec SOA. The interaction between diagnosis and SOA was not significant.
Mean percentages of correct responses by group and time interval
Figure 2. Mean percentage of correct letter identifications for the patients with migraine with aura (▪) and control subjects (▴) at each time interval. Error bars denote standard errors of the mean.
An alternative method of examining accuracy data was also used, in which suppression was defined as the number of trials out of nine at each SOA in which one or fewer letters were correctly identified. This alternative method of scoring gave rise to similar results: overall, MA patients were significantly more accurate than C subjects (F(1,12) = 5.07; p = 0.04).
Although these analyses showed a clear superiority in performance for the MA patients overall, the method of comparing group means has been criticized for not providing a comprehensive picture of performance in situations such as these.30 In particular, the variance in individual performance is obscured so that the means of the groups as a whole may not reflect accurately the data of any one given individual in that group. Additional analyses were therefore conducted. The mean PC at extreme SOA (40 and 190 msec), when it was not expected that suppression would occur, were compared between the groups. There was no difference in performance of the MA patients and the C subjects at either SOA. In order to obtain a measure of the spread of suppression, the area of the curve under a criterion level of performance was compared between the groups.30 There are several ways in which such a criterion might be defined. As it is complex to define chance levels of performance with letter stimuli, we initially followed the definition used in the threshold determination phase of the experiment. Here, suppression was deemed to have occurred if one or fewer letters (≤33% accuracy) were correctly reported from the given stimulus, whereas a correct report of two or more letters (≥67% accuracy) was judged as no suppression. There is hence an indeterminacy between these two extremes, the midpoint of which is 50%. It initially seemed appropriate to select this as the criterion. However, only one of seven MA patients had a curve that dropped below 50% at any SOA, and therefore between-group statistical comparisons were impossible. Hence, we redefined a cut-off performance level of 67% accuracy (suppression of only one letter when considering a single trial). The mean area of the curve below 67% in the MA patients was 353.29 arbitrary units (SD 330.5) and for the C subjects 1069.08 (SD 648.15). This difference was significant (t(12) = 2.6, p < 0.05). Inspection of individual plots (figures 3 and 4⇓) suggests that the significant difference between the groups arises because the time window during which suppression was possible in the MA patients was significantly narrower than in the control subjects.
Figure 3. Individual suppression plots for control subjects (percentage of correct letter identifications versus time interval). Here and in figure 4, individual plot symbols are omitted for clarity. The dashed horizontal line denotes the criterion level of performance for the area-under-curve analyses.
Figure 4. Individual suppression plots for patients with migraine.
Discussion.
Our findings confirm that visual stimulus processing is disturbed in migraine with aura to the extent that it is more difficult in these patients, as compared to matched controls, to suppress perception of simple targets by a TMS pulse over the primary visual cortex. Several points of interest concerning patient/control differences emerge from the study results. First, there was a significant difference in perceptual accuracy at the 100 msec interval, at which the migraine subjects showed a clear superiority in target visibility compared with controls. This observation matches very well with the results we obtained using metacontrast masking.16 Moreover, our study results are highly consistent with an earlier finding of superiority of target recognition in migraineurs.4 Second, an area-under-curve analysis demonstrated that a significant feature of the performance of MA subjects was the narrower time window (smaller bandwidth) in which suppression occurred.
In 1994, we speculated about a possible role of selective damage to intracortical inhibitory neurons in the primary visual cortex of migraineurs as a major factor in hyperexcitability and the origin of spontaneous cortical spreading depression (CSD).31 In human motor cortex, it has been shown that the cortical stimulation silent period (CSSP), a measure of local cortical inhibition, is shortened in migraine, possibly pointing to a deficiency of cortical inhibition.32 Our recent findings demonstrating that metacontrast masking is reduced in migraine with aura are also consistent with reduced cortical inhibition.16 The current study employed a parallel experimental procedure in which the target consisted of a letter trigram and the suppressing stimulus was a transcranial magnetic stimulus over V1 (rather than the spatially nonoverlapping visual mask used in the metacontrast technique). The response curves (stimulus recognition versus target-pulse-interval) in the two designs are remarkably similar, with maximal suppression of perception occurring when the visual mask/TMS pulse is delivered approximately 100 msec after target onset. Although it has been suggested that TMS suppression of letter targets may occur at earlier intervals,27 these findings were much less consistent across subjects (and actually absent in some) and generally weaker; we therefore concentrated on the suppression period centered on 100 msec poststimulus as being the most reliable index for a comparison of patient and control groups.
Several mechanisms have been forwarded in an attempt to explain the neuronal correlates of TMS-induced visual suppression. First, it can be argued that the masking effect of TMS might be caused by a collision of incoming visual impulses and the excitation of thalamo-cortical projections.22 However, as phosphenes were not perceived by subjects when TMS was applied without a visual stimulus, it has been noted that visual suppression cannot be regarded as masking by such an excitatory effect.20 Second, evidence suggests that the visual cortex may be suppressed either by the local immediate interference of the induced electric field with neuronal membrane potential and resulting inhibition of action potentials20 or through the generation of cortical inhibitory postsynaptic potentials (IPSP),21,22⇓ possibly due to TMS-induced synchronization.23 It is thus possible, by direct analogy with the explanation of metacontrast masking presented in the introduction to this article, that the cortical IPSP generated by the magnetic pulse serve to inhibit activity in cells processing information about the letter targets, thereby attenuating their visibility.
Given the foregoing, it is a plausible and reasonable claim that the enhanced visual accuracy at the 100 msec interval demonstrated in this study represents reduced cortical inhibition in migraineurs. Further support for this argument comes from the finding of a recent pilot study conducted in our laboratories that visual suppression is augmented in migraine with aura subjects after 15 minutes of 1-Hz repetitive magnetic stimulation (rTMS) of the visual cortex. rTMS has recently been reported to increase the inhibitory potential of cortical networks.33
With regard to the narrower time window in which suppression occurred in the patients with migraine, it is interesting to consider possible mechanisms. It would appear that the timing of the magnetic pulse is more critical in the migraineurs in order to interfere with the processing of the visual stimulus. One way of interpreting this finding is that it may be due to synchronization of impulses before or at the level of V1. It remains speculative to predict what this possible hypersynchronization of signal processing in the visual system of migraineurs with aura implies at the level of neurophysiological processes in neuronal circuitry.
One possible criticism of the current results is that they were obtained from a rather small number of subjects; however, it must be emphasized that the testing procedure is objective, and under computerized control. Furthermore, a large number of data points are obtained from each individual. Nonetheless, it remains possible that the relatively large within-group variances might have obscured statistical effects: clearly, a replication of the findings we report here will be desirable.
A second possible criticism is that, although suppression in the second phase of the study was markedly different between groups at the 100 msec SOA, thresholds for suppression (at the same SOA) in the first phase were not significantly different. This could give the impression that the results obtained in the suppression experiment were simply an artifact of inadequate threshold estimation. Although a bias in estimating thresholds cannot be ruled out in an unblinded design, we do not believe that threshold estimation errors can account for the threshold-suppression paradox, considering the stringent method of fine-tuning we applied. We believe that it is more likely that this apparent contradiction was due to the relatively small number of threshold determination trials, which would certainly have resulted in a very considerable variance in levels of suppression. One must thus be extremely cautious in inferring that the results in phase one contradict the data of the suppression phase. In this respect, there is no doubt that it would have been more informative to obtain robust stimulus-response curves at the 100 msec suppression interval. However, to obtain meaningful results, this would have implied many more magnetic stimuli, considerably prolonging the duration of each experiment. We therefore chose not to use such a design, but rather apply the randomized time-course experiment.
Enhanced responsiveness of the visual cortex is now a widely appreciated characteristic of the migrainous brain, manifested as an increased sensitivity to various physiologic environmental stimuli,34-36⇓⇓ deficient habituation, and increased excitability of the visual cortex.5,6,13⇓⇓ Magnetic stimulation of the occipital cortex has shown fairly consistently that the thresholds for the induction of phosphenes are lowered in migraine,8-11⇓⇓⇓ but a reliable interpretation of the data from the various studies may have been hampered by differences in methodology and the subjective nature of the parameter under consideration.18 The increased objectivity of the suppression of perception technique reported in this article has considerable potential, we believe, in advancing the state of knowledge about cortical dysfunction in patients with migraine, in particular by substantiating or negating these earlier claims of enhanced excitability. Use of the same assessment techniques before and after the introduction of a pharmaceutical agent with a known neuromodulatory effect would also add to its potential for aiding the understanding of the pattern of normal and dysfunctional neurophysiologic processes in the migrainous cortex.
These results obtained in this case-control sample confirm the previously reported hyperexcitability of the occipital cortex in migraine with aura,8 and provide preliminary although highly consistent and converging evidence that it may be due to attenuated cortical inhibition.
Acknowledgments
Supported by Project Grant 115 from The Migraine Trust.
Acknowledgment
The authors are grateful to Dr. Robin Henderson, of the Medical Statistics Unit, Lancaster University, for invaluable advice on area-under-curve analyses.
- Received May 3, 2000.
- Accepted October 3, 2000.
References
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Topics Discussed
Alert Me
Recommended articles
Transcranial magnetic stimulation confirms hyperexcitability of occipital cortex in migraine
Visual cortex hyperexcitability in migraine in response to sound-induced flash illusions
Thickening in the somatosensory cortex of patients with migraine
Decreased number and function of endothelial progenitor cells in patients with migraine