Abstract
We have reported a preliminary study confirming hyperexicitability of occipital cortex in migraine with aura (MwA) using transcranial magnetic stimulation (TMS). We have now completed a blinded study to investigate the occipital cortex in MwA and without aura (MwoA) compared with normal controls (NC) using TMS. TMS was performed using the Caldwell MES-10 stimulator. A circular coil 9.5 cm diameter was applied to the occipital scalp (7 cm above the inion). Stimulator intensity was increased in 10% increments until subjects reported visual phenomena or 100% intensity was reached. Stimulation intensity was then fine tuned to determine the threshold at which phosphenes were seen. Fisher's exact t-test and logrank test were used for statistical comparisons. Ten subjects with MwA and MwoA were compared to 10 NC. The difference in the proportion of subjects with phosphene generation was statistically significant (MwA 100%, MwoA 60% and NC 30%) [P = 0.003]. The difference in threshold levels for phosphenes was also significant for MwA 42.8%, and controls 57.3% [P = 0.0001]. There is a difference in threshold for excitability of occipital cortex in MwA and MwoA compared to NC. This is a direct neurophysiological correlate for clinical observations, which have inferred hyperexicitability of the occipital cortex in migraineurs.
Introduction
Enhanced excitability of occipital cortex neurones has been proposed as the basis for the spontaneous or triggered onset of the migraine aura (1). Clinical symptoms also suggest that the occipital cortex have an important role in the elaboration of a migraine attack. In a large cross-sectional study, 90% of subjects with migraine with aura had visual disturbance (2). The visual phenomena commonly reported include negative symptoms such as blind spots or scotomas and positive phenomena such as fortification spectra and photopsias. Clinical analysis of the neurologic features of visual aura (3) suggests a wave of intense excitation propagates at a rate of approximately 3 mm per minute across the visual cortex, in a manner resembling a neuroelectric property of certain animal brains that is termed spreading depression (SD) of Leao (4).
Recent clinical studies of visual discrimination demonstrated that migraine sufferers had a greater sensitivity for low-level visual processing between attacks (5, 6). However, neurophysiologic evidence for hyperexcitability of occipital cortex between attacks is both limited and controversial. For instance, some studies have demonstrated differences in the amplitude of the visual evoked responses in migraineurs compared to controls (7–9), while others found no abnormalities (10, 11). Transcranial magnetic stimulation (TMS), however, has provided a new opportunity to noninvasively and directly investigates occipital cortex physiology and excitability (12–14). When the motor cortex was studied, three studies reported a different threshold for motor evoked potentials in migraineurs compared to normal controls (15–17). We had reported a preliminary study confirming hyperexcitability of the occipital cortex in migraine with aura in the interictal period (18). Later, a study was reported with opposite results (19), however, recently 3 others have confirmed our initial finding (20–22). We now report a blinded study in which both, migraine with and without aura was compared to controls for the threshold of magnetophosphenes. A blinded study is very important since phosphene visualization is a subjective phenomena.
Methods
Clinical subjects
Ten subjects with the diagnosis of migraine with aura (MwA) mean age 38 ± 13 (9 females, 1 male), 10 migraine without aura (MwoA) mean age 39 ± 10 (8 females, 2 males) were compared to 10 normal controls mean age of 37 ± 9 (8 females, 2 males). Independent personnel selected the subjects by reviewing charts from the headache clinic and all subjects were different from the previous report (18). The diagnosis was made according to the International Headache Society (IHS) criteria (23). For MwA, subjects who had visual aura and in particular a propensity towards visual triggers were recruited. The control subjects were recruited from hospital personnel and a full headache history was recorded. Subjects who had more than one muscular contraction headache per month, history of seizures, pacemakers and those on drugs known to alter central nervous system excitability were excluded. The Institutional Review Board on human experimentation of the Henry Ford Hospital approved the protocol. All subjects gave fully informed consent with regards to the procedure and risks of transcranial magnetic stimulation were explained to the subject.
All study participants had a normal neurological examination at the time of study. In migraine subjects the study was performed in the interictal period with a headache free interval of at least one week. The investigator was blinded to the diagnosis both during the collection and analysis of data. The subjects were weaned off preventive drugs four weeks prior to the study. None of the subjects were taking drugs known to alter central nervous system excitability (sedatives, hypnotics, anticonvulsants or β-blockers; 24) for at least four weeks prior to the study. Analgesic and abortive medications were permitted during this four-week period. None, however, were taken in the 48-h period prior to the study.
Technique: All eligible subjects underwent occipital cortex stimulation using the Cadwell Magstim apparatus (Cadwell Laboratories Inc. Kennewick, WA, USA). A 9.5-cm circular coil was used which has 14 turns giving a peak magnetic field strength of 2 Tesla and 530 V/m of peak electric field strength. The recordings were conducted in a semidarkened room. Subjects were asked to wear a blindfold, and to close their eyes to diminish any ambient light. Subjects in each group were blindfolded for five minutes or less only during the experiment. There was no light deprivation beyond 10 min since after 45 min this has been shown to alter phosphene threshold (24).
Measurements were conducted according to the following protocol. The inion was palpated and a line joining nasion to inion was made. The rim of the 9.5-cm diameter circular coil was then applied to the occipital scalp in the midline, 7-cm superior to the inion. The magnetic coil was applied tangentially to reduce excessive contractions of the neck muscles. To avoid bias subjects were not informed of what to expect or which area of the brain was being stimulated, but were asked to report all sensory experiences during stimulation, including muscle twitches and sensations of vision, smell, or taste. Next, stimulator intensity (SI) was increased in steps of 10% (percentage refers to fraction of maximal stimulation output) until the subjects either reported seeing phosphenes (bright scintillations in the subjects visual field) or until a maximum of 100% intensity of stimulation was reached. (Maximal stimulation intensity is technically built in to the equipment). Stimulation intensity was then fine tuned to determine the threshold at which phosphenes could just be visualized in those subjects reporting this experience. The optimum position of the coil was determined where phosphenes were visualized clearly at lower SI. In those subjects who did not report seeing phosphenes when 100% stimulation intensity was achieved, the stimulator was then moved in 1 cm steps to define a different point for stimulation and the procedure repeated. For stimulation of either hemisphere, different coil positions were used. Clock-wise coil currents were used for excitation of the left visual cortex and counter-clockwise for the right. The frequency at which stimulations were delivered was 1 every 20 s or 5/min. The lowest SI where phosphenes was generated was termed phosphene threshold (PT) (18) and recorded for migraine and controls.
Statistical analysis
Fisher's exact t-test was used to compare the proportion of subjects reporting phosphenes between MwA and control group. A logrank test was then done to compare the threshold levels for phosphene generation between the three groups. The investigator remained blinded to the diagnosis till the data analysis was complete.
Results
All 10 MwA patients, six of 10 MwoA reported phosphenes during occipital cortex stimulation whereas only three of 10 controls visualized phosphenes. The visual phenomena were reported as white spots in the visual fields, which were present only momentarily and did not last beyond the stimulation. When the coil was in the midline and 7-cm superior to the inion, the phosphenes were visualized in the inferior half of both visual fields. The difference in the proportion of subjects with phosphene generation between migraine and control groups was significant (MwA 100%, MwoA 60% and controls 30, P = 0.003, Fisher's exact test) (Fig. 1). Pairwise comparisons of the three groups showed that the difference between MwA and controls was statistically significant (P = 0.003, Fisher's exact test). The difference between MwA and MwoA was of borderline statistical significance (P = 0.035, Fisher's exact test). There was no difference between MwoA and controls. The overall difference in threshold levels among the three groups was statistically significant (MwA 42.8 ± 11.4%, MwoA 55.7 ± 12, controls 57.3 ± 23.9%, P = 0.0001, logrank test) (Figs 1 and 2). Pairwise comparisons of the three groups showed that the differences between MwA and the other two groups were also statistically significant (controls P = 0.003 and MwoA 0.007, logrank tests). The difference between MwoA and controls was not statistically significant. All threshold levels for MwA were lower than the lowest level in the control subjects. Interestingly, one of the controls reported to the neurology clinic the next day with visual blurring. She gave a history of similar phenomena, occurring approximately twice a year for the last year with occasional headache associated with these. On reviewing the history another neurologist diagnosed the patient as migraine aura without headache. This particular subject had not admitted to a history of migraines or other headache, during the screening and hence inadvertently been enrolled as a normal control. Following this, the individual recruiting the patients was alerted and a questionnaire was sent again to all the normal controls. One of the other three controls who visualized phosphenes was also, diagnosed as migraine without aura. The third control that visualized phosphenes on TMS had a family history consistent with migraine with aura in his mother and maternal grandmother but the subject had no history of migraines.
Phosphene proportion and threshold. ▪ MwA; MwoA; □ controls.
Comparison of threshold levels for MwA, no aura and control subjects. ······· MwA; ––––– MwoA; ——— controls. P = 0.0001 (Logrank test).
Discussion
Our results have confirmed that the threshold for observing phosphenes induced by TMS in migraine with aura subjects is lower than in controls. In the last two years several other centres (20–22) have confirmed this finding but this is the only study to the best of our knowledge that assessed phosphenes in a blinded manner. Thus there is an increasing body of evidence that the excitability threshold of the occipital cortex in migraine sufferers is low compared to normal, thus strongly indicating that the occipital cortex neurones can be hyperexcitable in this condition. The migraine without aura group also had proportionately more phosphenes than the controls similar to a recent report of a reduced phosphene threshold in mwoA. In our study however, the threshold for phosphene generation was only significant once the subjects who were inappropriately classified as controls were removed. Thus, there seems to be a spectrum for the generation of phosphenes by TMS where MwA has the lowest threshold, controls the highest and MwoA lies somewhere in between.
A limitation of our study is the measurement of sensory phenomena of phosphenes, which lacks objective record. This potential was overcome in part with the investigator being blinded to diagnosis during testing. Also, neither the migraine or control subjects were alerted to the nature of the potential sensory phenomena they might experience. Despite this, and appropriate to our stimulation locus, the sensory symptoms complained of were solely visual. Further, phosphenes were only experienced in the appropriate visual field at a time when the patients had no knowledge of the coil position or, presumably, no knowledge of neuroanatomy. The subjects also reported phosphenes as being brighter and more rapidly saturated at higher stimulation intensities, again, of which they had no knowledge. The inadvertent recruitment of two migraine patients as controls could be considered a limitation of the study. However, we felt recruiting additional controls after breaking the blind would have changed the study design in its entirety. We therefore chose to maintain the a priori study design similar to ‘intent to treat’ statistical design. Further, if we had included the two controls in the migraine group this would have skewed the results in favour of the hypothesis.
We also believe that inclusion of these controls is the result of migraine being an under diagnosed condition and therefore the two subjects on initial screening denied troublesome headaches (25). This in fact, may be why there has been a lot of variance in different studies for phosphene generation and threshold in ‘normal controls’. A recent study where ‘normal controls’ were asked about a personal history of migraines single pulse TMS did not generate phosphenes in a majority of those subjects (26). Therefore, the controls in our present study may be outliers in the physiological spectrum of phosphene generation.
Our results are similar to most studies reported recently (20–22) and only opposite to the one study reported by Afra et al. (19). In their study Afra et al. (19) found no differences in proportion of subjects who had phosphenes in migraine and healthy volunteers. Further, the threshold for phosphenes in controls was lower than migraine and therefore opposite to all other reports thus far (18, 20–22). The differences in controls may be explained on the basis of methodology. In our study we used the Caldwell MES-10 stimulator while Afra et al. (19) used the Magstim-200. In earlier studies where normal subjects were studied using the Caldwell MES-10, the investigators had immense difficulty in eliciting phosphenes (12). The results in our controls are similar to those reported by others using the same stimulator and coil size (12). Recently, with the use of the different magstim devices, i.e. Magstim-200 some investigators have been able to demonstrate phosphenes in ‘normal’ subjects (27–29). However careful review of the studies performed in normal controls where the proportion of phospenes were 67% (27), 68% (28) and 82% (29) showed no screening for migraine as part of the exclusion criteria. The proportion of phosphenes in the control group however, is higher than ours in the studies where migraineurs were excluded, 94% (21), 89% (19). The coil size is different, in our study a 9.5-cm coil was placed tangentially over the occiput with the focal point 7-cm superior to the inion. Afra et al. (19) used a larger coil and therefore, theoretically generated more powerful electric current resulting in greater probability for generation of phosphenes. The third difference was coil positioning, and with recent studies emphasizing coil position in the motor cortex (30), this could also be another reason why there were different results obtained. In our study the coil was held tangentially over the occiput with the handle inferior. It is uncertain if coil positioning has relevance for the occipital cortex since studies have been performed in the motor cortex however, phopshenes are most likely a neuronal event and therefore, coil position may be important. We do not believe light deprivation could have accounted for differences in these two studies since our subjects were blind-folded for 10 min at the most and difference in phosphene threshold were only noted to be significant after 45 min of light deprivation.
The lack of phosphene generation in the migraine group (19) however, is perplexing particularly in light of more recent studies (20–22) but the reason for this may be two-fold. First there were technical differences in the study as stated above. At higher levels of stimulus intensities (SI) phosphenes tend to saturate and disappear (29) and therefore theoretically the SI for the migraine group may have reached threshold effect. The second reason may be due to subject selection, in our study only subjects who had visual aura and especially whose headaches were triggered by visual stimulation were recruited. In the study by Afra et al. (19) the subjects there was no mention whether the patients had predominant visual aura and further, those some patients with migraine with aura did not have aura in 20% of their attacks.
That the occipital cortex neurones are in a state of enhanced excitability lends support to the theory of ‘central neuronal hyperexcitability’ being a major factor in the cause of aura, the evidence for which has been reviewed (1) and noted earlier in this paper. The concept is entirely compatible with a heightened susceptibility to spontaneous or triggered synchronous depolarization of occipital cortex neurones that initiates a spreading depression-like event, currently favoured as the basis of the migraine aura (31). Migraine with and without aura may share similar pathophysiology, which has been exemplified by a PET study (32). Since our study has shown a range of occipital cortex excitability being highest in MwA and lowest in normal controls with MwoA somewhere in between, this again is further clue of similar pathogenic mechanisms for these disorders.