Somatosensory evoked high-frequency oscillations reflecting thalamo-co
http://www.100md.com
《大脑学杂志》
1 Headache Research Unit, University Departments of Neurology
2 Neuroanatomy, University of Liege, CHR Citadelle, Liege, Belgium
3 Headache Clinic, INM Neuromed, Pozzilli (IS), Italy
Summary
A deficit of habituation in cortical information processing, including somatosensory evoked potentials (SSEPs), is the most consistent neurophysiological abnormality in migraine patients between attacks. To explore further the mechanisms underlying this interictal neural dysfunction, we have studied the high-frequency oscillations (HFOs) embedded in SSEPs because they are thought to reflect spike activity in thalamo-cortical cholinergic fibres (early HFOs) and in cortical inhibitory GABAergic interneurons (late HFOs). Untreated migraine patients with (MA) and without (MO) aura were recorded during (n = 13: nine MO, four MA) and between attacks (n = 29: 14 MO, 15 MA) and compared with healthy volunteers. SSEPs were filtered off-line (digital band-pass between 450 and 750 Hz) to extract the two HFO bursts from the broad-band contralateral N20 somatosensory cortical response obtained by median nerve stimulation. In both migraine groups, amplitudes and latencies of conventional broad-band SSEPs recorded interictally from cervical and parietal active electrodes were not significantly different from those found in healthy volunteers. In contrast, maximum peak-to-peak amplitude and area under the rectified curve of the early HFO burst were significantly smaller in both MA and MO patients than in healthy volunteers. There was no significant difference in the later HFO burst between migraineurs and healthy volunteers. During attacks, all electrophysiological measurements in migraineurs were similar to those found in healthy volunteers. Thalamo-cortical activation, as reflected by the early SSEP HFO burst, may thus be reduced in migraine interictally, but normalizes during an attack, whereas intracortical inhibition, as indexed by the late HFO burst, is normal at any time. This supports the hypothesis that the habituation deficit in migraineurs is due to a reduced pre-activation level of sensory cortices and not to increased cortical excitability or reduced intracortical inhibition.
Key Words: high-frequency oscillations; somatosensory evoked potentials; habituation; migraine; cortical excitability
Abbreviations: HFO = high-frequency oscillation; MA = migraine with aura; MO = migraine without aura; SSEP = somatosensory evoked potential
Introduction
The best reproducible interictal electrophysiological abnormality in migraine patients with (MA) or without aura (MO) is lack of habituation (or even potentiation) of evoked and event-related cortical potentials during repetition of the same stimuli (see review by Schoenen et al., 2003). Such a habituation deficit was also reported for somatosensory evoked potentials (SSEPs; Ozkul and Uckardes, 2002). It is most probably responsible for the strong intensity dependence of auditory evoked cortical responses (IDAP) found in migraineurs between attacks (Ambrosini et al., 2003) and for the increased amplitude of evoked potentials reported in some studies (Gawel et al., 1983; Schoenen et al., 2003). Moreover, it has a familial character (Sandor et al., 2002).
The underlying cause of the habituation deficit in cortical information processing is still under debate. In theory, it could be due to increased cortical excitability (and homosynaptic facilitation) or to decreased intracortical inhibition (heterosynaptic inhibition) (Stanley, 1976), both of which have been proposed in migraine following transcranial magnetic stimulation (Aurora et al., 1998; Aggugia et al., 1999; Mulleners et al., 2001; Battelli et al., 2002; Young et al., 2004) and psychophysical studies of the visual cortex (Chronicle and Mulleners, 1994; Wray et al., 1995). Conversely, both methods have also provided results favouring hypoexcitability of the visual cortex and normal function of inhibitory cortical interneurons (afra et al., 1998; Shepherd, 2001; Shepherd et al., 2001; Bohotin et al., 2002, 2003). We have therefore hypothesized, in line with the ‘ceiling theory’ (Knott and Irwin, 1973), that an interictal reduction of the pre-activation level of sensory cortices could be the culprit in migraine. It could indeed explain the low amplitudes found in the first blocks of evoked responses and the lack of habituation in subsequent blocks since it would offer a larger range for suprathreshold cortical activation up to the ‘ceiling’ where habituation starts (Schoenen, 1996). The pre-activation excitability level of sensory cortices is set by activity in thalamo-cortical loops and aminergic, in particular serotoninergic and noradrenergic, projections from the upper brainstem (Mesulam, 1990; Hegerl and Juckel, 1993).
Using digital high-pass filtering (>400 Hz), it was shown in recent years that high frequency (±600 Hz) oscillations (HFOs) of low amplitude are superimposed on the N20 component of median nerve SSEPs. Two bursts of HFOs have been identified in animals (Ikeda et al., 2002; Baker et al., 2003), and with depth (Klostermann et al., 1998, 1999) as well as scalp recordings in humans (Gobbele et al., 1998, 1999, 2003; Hashimoto et al., 1999; Halboni et al., 2000; Haueisen et al., 2001). Early HFOs are thought to be generated by thalamo-cortical afferents, and late HFOs by inhibitory interneurons in parietal area 3b. A study of SSEP HFOs in migraine may thus shed some light on the neurophysiological causes of the interictal habituation deficit found in evoked cortical responses, as it explores simultaneously thalamo-cortical activation and intrinsic inhibition of the parietal cortex. We reasoned that on the one hand the early burst of HFOs would be increased if migraineurs have cortical hyperexcitability between attacks, but decreased if thalamo-cortical pre-activation is reduced. On the other hand, defective intracortical inhibition would result in a decrease of late HFOs.
Material and methods
Subjects
We recorded 42 migraine patients: 19 MA [International Classification of Headache Disorders, II (ICHD-II), Headache Classification Committee of the International Headache Soiety (2004) code 1.2; mean age: 31 ± 11 years.; 11 women, eight men] and 23 MO (ICHD-II code 1.1; mean age: 30.0 ± 10; 17 women and six men). They were compared with 15 healthy volunteers of comparable age and gender distribution (healthy volunteers; mean age 29 ± 7; eight women and seven men) without personal or familial history of migraine and devoid of any detectable medical condition. Subjects taking any medication on a regular basis except for the contraceptive pill were excluded.
We analysed separately the recordings from two groups of patients. In the first (interictal) group (n = 29; 15 MA, 14 MO, 17 females, 12 males), we included only data from patients who had an interval of at least 3 days between the recording and their last or their next migraine attack. The second group (n = 13; 4 MA, 9 MO, 11 females, two males) was composed of subjects in whom the recordings were made within a time range of 12 h before or after the beginning of an attack.
All participants gave informed consent in accordance with the Declaration of Helsinki and the study was approved by the Ethics Committee of the Faculty of Medicine, University of Liege.
Data acquisition
SSEPs were elicited after electrical stimulation of the right median nerve at the wrist using a constant current square wave pulse (0.2 ms width, cathode proximal) and a stimulus intensity set at twice the motor threshold. One series of 500 sweeps was collected and averaged at a repetition rate of 4.4 Hz. The active electrodes were placed over the contralateral parietal area (C3', 2 cm posterior to C3 in the international 10–20 system) and on the fifth cervical spinous process (Cv5), both referenced to Fz; the ground electrode was on the right arm. The evoked potential signals were amplified by CEDTM 1902 pre-amplifiers and recorded by a CEDTM 1401 device (Cambridge Electronic Design Ltd, Cambridge, UK). Subjects were seated relaxed on a comfortable chair in an illuminated room and asked to stay with their eyes opened and to fix their attention on the wrist movement. Forty milliseconds of the post-stimulus period were sampled at 5000 Hz to collect a wide range of frequencies. All recordings were averaged off-line using the SignalTM software package version 2.14 (CED Ltd).
Data analysis
Wide-band somatosensory evoked potentials
For wide-band unfiltered SSEPs (0–2500 Hz), we have identified the various components according to their respective latencies. Thereafter we have measured peak-to-peak amplitudes of the N13 cervival component under the Cv5 active derivation and of N20onset–N20peak, N20peak–P25, P25–N33 under the active C3' left-sided scalp electrode.
High-frequency oscillations
Digital zero-phase shift band-pass filtering between 450 and 750 Hz (Barlett-Hanning window, 51 filter coefficients) was applied off-line in order to extract the HFOs superimposed on the N20 left parietal component of the SSEP. In the majority of recorded traces, we were able to identify two separate bursts of HFOs: the first, early burst occurred in the latency interval of the ascending slope of the conventional N20 component, and the second, late burst in the time interval of the descending slope of N20, sometimes extending into the ascending slope of the N33 peak. In general, the frequency of oscillations was higher in the first than in the second HFO burst, and in between the first and the second burst there was a clear frequency and amplitude decrease, which allowed separation of the two bursts. In recordings in which a clear distinction between the two components was not possible, we considered as first burst HFOs occurring before the N20 peak and as second burst those after the N20 peak.
After elimination of the stimulus artefact, we have measured the latency of the negative oscillatory maximum, the intra-burst frequency, the number of negative peaks, the maximum peak-to-peak amplitude and the area under the rectified curve (AUC), computed using an interval of ±1.17 ms around the maximum peak to provide stability against noise (see Gobbele et al., 1998). All measurements were performed separately on the two HFO bursts.
Finally, since the recorded signal was short (50 ms) and contained high-frequency oscillations that varied with time, we have used a short-time Fourier transform spectrum (STFT) to obtain an indication of the time–frequency properties. This method offers the advantage of visualizing the power (z-axes) of a two-dimensional signal represented by time (x-axes) and frequency (y-axes).
The difference between each group of migraine patients and healthy volunteers was evaluated by one-way analysis of variance (ANOVA), and a P < 0.05 was considered significant.
Results
All the measured components, for both wide-band and HFO SSEPs, were clearly identified in all examined subjects. Tables 1 and 2 show synoptically mean values ± 1 SD and levels of significance found for the various measures taken in healthy volunteers, in interictal recordings from MA and MO patients and in ictal recordings.
Interictal recordings
Broad-band SSEPs
In the MO and the MA groups, peak-to-peak amplitudes and latencies recorded from the cervical and parietal active electrodes were not significantly (P > 0.05) different from those of healthy volunteers.
High-frequency oscillations in SSEPs
Maximum peak-to-peak amplitude and area under the rectified curve of the first burst of HFOs were significantly smaller in MO [F(1,27) = 4.822, P = 0.037 and F(1,27) = 5.456, P = 0.027, respectively] and MA patients [F(1,28) = 7.358, P = 0.011 and F(1,28) = 7.067, P = 0.013] than in healthy volunteers.
In contrast, there was no significant difference between healthy volunteers and patients in amplitude or AUC of the second HFO burst.
Latency of the negative oscillatory maximum, intra-burst frequency and number of negative peaks for both early and late HFO bursts were similar between the three subject groups.
Ictal recordings
In the 13 migraineurs recorded during an attack or in the immediate peri-attack phase, measurements performed on both broad-band and HFOs SSEPs were not significantly different from those of our healthy volunteer group.
Discussion
In accordance with previous studies (Firenze et al., 1988; de Tommaso et al., 1997), we found no significant difference between migraineurs recorded outside of an attack and healthy volunteers for broad-band conventional SSEPs. This suggests that the somatosensory signal is normally conducted along the dorsal columns up to the contralateral somatosensory cortex. In contrast, after band-pass filtering, the first early burst of high-frequency oscillations embedded in the N20 component of the SSEP was significantly reduced in both MA and MO subjects compared with healthy volunteers, while the second late HFO burst was normal.
To the best of our knowledge, our study is the first to apply the rather recent methodology of SSEP HFOs to research on brain physiology in migraine patients. The results have to be interpreted in the light of available, still incomplete and partially contradictory knowledge on the neural generators involved in HFOs.
As mentioned, the late burst of HFOs is thought to have a postsynaptic cortical origin, probably reflecting the activity of intracortical GABAergic inhibitory interneurons (Hashimoto et al., 1996; Mochizuki et al., 2003). Presynaptic subcortical activity, most probably repetitive action potentials along the terminal segment of thalamocortical fibres, is on the contrary responsible for the early burst of HFOs (Gobbele et al., 1998, 2003; Ikeda et al., 2002; Mochizuki et al., 2003). Multichannel scalp recordings and dipole source analysis in normal subjects have shown that both the thalamic and the cortical components of the early presynaptic component of SSEP HFOs were reduced in amplitude during non-rapid eye movement (REM) sleep with partial recovery during REM sleep, while the later HFO bursts were unchanged or rather enhanced (Halboni et al., 2000). Furthermore, the amplitude of the thalamic component and of the cortical tangential source activity of HFOs was significantly greater with eyes open than with eyes closed (Gobbele et al., 2000). After administration of the central acetylcholinesterase inhibitor rivastigmine, Restuccia et al. (2003) found an increase in amplitude of the cortical HFO component in the 18–28 ms latency range and concluded that cortical HFOs are generated mainly by specialized pyramidal neurons. Considering that cholinergic systems play a pivotal role in sleep–wake cycle and arousal (Mesulam, 1995), cholinergic reticulothalamic pathways could be crucial for the generation of SSEP HFOs.
Our finding of a reduced activity of the early HFO burst in migraineurs between attacks suggests, therefore, that migraine is associated with hypofunction in thalamo-cortical excitatory cholinergic afferents. By the same token, the normal second HFO burst indicates that the late generators function normally, be they cortical inhibitory interneurons (Hashimoto et al., 1996; Mochizuki et al., 2003) or pyramidal chattering cells (Restuccia et al., 2003).
As mentioned in the Introduction, controversy has arisen about the excitability status of the cerebral cortex in migraineurs between attacks. Intuitively, some clinical features of migraine such as hypersensitivity to light or noise, and the aura symptoms suggest that the cerebral cortex is hyperexcitable. In the physiological sense, however, hyperexcitability would mean that the brain responds to a subliminal stimulus or that its response to a supraliminal stimulus is increased in amplitude. These two abnormalities have never been demonstrated precisely in migraineurs. Increased amplitudes of psycho-physical (Wray et al., 1995) or cortical evoked responses (Gawel et al., 1983) were reported by some, but these studies are based on the use of repetitive stimuli and we and others have shown that the lack of habituation of an initial low amplitude response during stimulus repetition is responsible for the net increase in averaged responses (Schoenen et al., 2003). Unfortunately, transcranial magnetic stimulation of the visual cortex in migraineurs has provided contradictory results (see the review by Schoenen et al., 2003).
We have therefore hypothesized that the finding of reduced habituation of evoked responses in migraine is probably not due to hyperexcitability and homosynaptic facilitation as such, nor to lack of heterosynaptic inhibition, but to a reduced pre-activation excitability level of sensory cortices (Schoenen, 1996). According to the ‘ceiling theory’ (Knott and Irvine, 1973), this can explain the low response amplitude for a single or a few stimuli and its marked increase (potentiation) or lack of decrease, i.e. habituation, because it allows a wider range of suprathreshold activation before reaching the ‘ceiling’ and initiating habituation of the response. Our finding of reduced early HFO bursts in SSEPs, and thus reduced thalamo-cortical activation is in agreement with this hypothesis and can hardly be explained by increased cortical excitability. Similarly, the normal late SSEP HFO suggests that intracortical inhibition is normal in migraine.
Interestingly, cortical evoked potentials habituate normally just before and during a migraine attack (afra et al., 2000) which suggests, according to the previous ‘ceiling’ hypothesis, that cortical excitability increases to normal levels. As expected, the early HFO burst was within normal values in migraineurs recorded during an attack, which probably reflects an increase in thalamo-cortical activation. Whether such an increase could be induced by the ictal activation of the dorsal rostral brainstem (Weiller et al., 1995; Bahra et al., 2001) which contains state-setting aminergic nuclei projecting to thalamus and cortex remains to be determined. Activation of these nuclei and of thalamo-cortical projections may contribute to an ictal normalization of the cortical pre-activation level.
Supplementary studies of HFO are needed and of interest in migraine. Comparative recordings in other primary headaches such as tension-type headache would be useful to verify the specificity of our findings for migraine. Studying the effect of pharmacological agents, among which prophylactic anti-migraine drugs and repetitive transcranial magnetic stimulations which are able to modify in opposite directions the excitability of the underlying cortex depending on stimulation frequency, can be used to modulate HFOs and shed light on the precise neuronal generators.
References
afra J, Mascia A, Gerard P, Maertens de Noordhout A, Schoenen J. Interictal cortical excitability in migraine: a study using transcranial magnetic stimulation of motor and visual cortices. Ann Neurol 1998; 44: 209–15.
afra J, Sandor PS, Schoenen J. Habituation of visual and intensity dependence of auditory evoked cortical potentials tend to normalize just before and during the migraine attack. Cephalalgia 2000; 20: 714–9.
Aggugia M, Zibeti M, Febbraro A, Mutani R. Transcranial magnetic stimulation in migraine with aura: further evidence of occipital cortex hyperexcitability. Cephalalgia 1999; 19: 465.
Ambrosini A, Rossi P, De Pasqua V, Pierelli F, Schoenen J. Lack of habituation causes high intensity dependence of auditory evoked cortical potentials in migraine. Brain 2003; 126: 2009–15.
Aurora SK, Ahmad BK, Welch KMA, Bdardhwaj P, Ramadan NM. Transcranial magnetic stimulation confirms hyperexcitability of occipital cortex in migraine. Neurology 1998; 50: 1111–4.
Bahra A, Matharu MS, Buchel C, Frackowiak RS, Goadsby PJ. Brainstem activation specific to migraine headache. Lancet 2001; 357: 1016–7.
Baker SN, Curio G, Lemon RN. EEG oscillations at 600 Hz are macroscopic markers for cortical spike bursts. J Physiol 2003; 550: 529–34.
Battelli L, Black KR, Wray SH. Transcranial magnetic stimulation of visual area V5 in migraine. Neurology 2002; 58: 1066–9.
Bohotin V, Fumal A, Vandenheede M, Bohotin C, Schoenen J. Excitability of visual V1–V2 and motor cortices to single transcranial magnetic stimuli in migraine: a reappraisal using a figure-of-eight coil. Cephalalgia 2003; 23: 264–70.
Bohotin V, Fumal A, Vandenheede M, Gerard P, Bohotin C, Maertens de Noordhout A, et al. Effects of repetitive transcranial magnetic stimulation on visual evoked potentials in migraine. Brain 2002; 125: 912–22.
Chronicle E, Mulleners W. Might migraine damage the brain Cephalalgia 1994; 14: 415–8.
de Tommaso M, Sciruicchio V, Tota P, Megna M, Guido M, Genco S, et al. Somatosensory evoked potentials in migraine. Funct Neurol 1997; 12: 77–82.
Firenze C, Del Gatto F, Mazzotta G, Gallai V. Somatosensory-evoked potential study in headache patients. Cephalalgia 1988; 8: 157–62.
Gawel M, Connolly JF, Clifford Rose FC. Migraine patients exhibit abnormalities in the visual evoked potential. Headache 1983; 23: 49–52.
Gobbele R, Buchner H, Curio G. High-frequency (600 Hz) SEP activities originating in the subcortical and cortical human somatosensory system. EEG Clin Neurophysiol 1998; 108: 182–9.
Gobbele R, Buchner H, Scherg M, Curio G. Stability of high-frequency (600 Hz) components in human somatosensory evoked potentials under variation of stimulus rate—evidence for a thalamic origin. Clin Neurophysiol 1999; 10: 1659–63.
Gobbele R, Waberski TD, Kuelkens S, Sturm W, Curio G, Buchner H. Thalamic and cortical high-frequency (600 Hz) somatosensory-evoked potential (SEP) components are modulated by slight arousal changes in awake subjects. Exp Brain Res 2000; 133: 506–13.
Gobbele R, Waberski TD, Thyerlei D, Thissen M, Darvas F, Klostermann F, et al. Functional dissociation of a subcortical and cortical component of high-frequency oscillations in human somatosensory evoked potentials by motor interference. Neuroscience 2003; 350: 97–100.
Halboni P, Kaminski R, Gobbele R, Züchner S, Waberski TD, Herrmann CS, et al. Sleep stage dependent changes of the high-frequency part of the somatosensory evoked potentials at the thalamus and cortex. Clin Neurophysiol 2000; 111: 2277–84.
Hashimoto I, Mashiko T, Imada T. Somatic evoked high-frequency magnetic oscillations reflect activity of inhibitory interneurons in the human somatosensory cortex. EEG Clin Neurophysiol 1996; 100: 189–203.
Hashimoto I, Kimura T, Fukushima T, Iguchi Y, Saito Y, Terasaki O, et al. Reciprocal modulation of somatosensory evoked N20 m primary response and high-frequency oscillations by interference stimulation. Clin Neurophysiol 1999; 110: 1445–51.
Haueisen J, Schack B, Meier T, Curio G, Okada Y. Multiplicity in the high-frequency signals during the short-latency somatosensory evoked cortical activity in humans. Clin Neurophysiol 2001; 112: 1316–25.
Headache Classification Committee of the International Headache Society. International classification of headache disorders, II. Cephalalgia 2004; 24 (Suppl 1).
Hegerl U, Juckel G. Intensity dependence of auditory evoked potentials as an indicator of central serotonergic neurotransmission: a new hypothesis. Biol Psychiatry 1993; 33: 173–87.
Ikeda H, Leyba L, Bartolo A, Wang Y, Okada YC. Synchronized spikes of thalamocortical axonal terminals and cortical neurons are detectable outside the pig brain with MEG. J Neurophysiol 2002; 87: 626–30.
Klostermann F, Nolte G, Losch F, Curio G. Differential recruitment of high frequency wavelets (600 Hz) and primary cortical response (N20) in human median nerve somatosensory evoked potentials. Neurosci Lett 1998; 256: 101–4.
Klostermann F, Funk T, Vesper J, Curio G. Spatiotemporal characteristics of human intrathalamic high-frequency (>400 Hz) SEP components. Neuroreport 1999; 10: 3627–31.
Knott JR, Irwin DA. Anxiety, stress and the contingent negative variation. Arch Gen Psychiatry 1973; 29: 538–41.
Mesulam MM. Large-scale neurocognitive networks and distributed processing for attention, language and memory. Ann Neurol 1990; 28: 597–613.
Mesulam MM. Cholinergic pathways and the ascending reticular activating system of the human brain. Ann NY Acad Sci 1995; 9 (Suppl 2): 23–8.
Mochizuchi H, Machii K, Terao Y, Furubayashi T, Hanajima R, Enomoto H, et al. Recovery function of and effects of hyperventilation on somatosensory evoked high-frequency oscillation in Parkinson's disease and myoclonus epilepsy. Neurosci Res 2003; 46: 485–92.
Mulleners WM, Chronicle EP, Palmer JE, Koehler PJ, Vredeveld JW. Visual cortex excitability in migraine with and without aura. Headache 2001; 41: 565–72.
Ozkul Y, Uckardes A. Median nerve somatosensory evoked potentials in migraine. Eur J Neurol 2002; 9: 227–32.
Restuccia D, Della Marca G, Valeriani M, Rubino M, Paciello N, Vollono C, et al. Influence of cholinergic circuitries in generation of high-frequency somatosensory evoked potentials. Clin Neurophysiol 2003; 114: 1538–48.
Sandor P, Ambrosini A, Agosti RM, Schoenen J. Genetics of migraine: possible links to neurophysiological abnormalities. Headache 2002; 42: 365–77.
Schoenen J. Deficient habituation of evoked cortical potentials in migraine: a link between brain biology, behavior and trigeminovascular activation Biomed Pharmacother 1996; 50: 71–8.
Schoenen J, Ambrosini A, Sandor PS, Maertens de Noordhout A. Evoked potentials and transcranial magnetic stimulation in migraine: published data and viewpoint on their pathophysiologic significance. Clin Neurophysiol 2003; 114: 955–72.
Shepherd AJ. Increased visual after-effects following pattern adaptation in migraine: a lack of intracortical excitation Brain 2001; 124: 2310–8.
Shepherd AJ, Palmer JE, Davis G. Increased visual after-effects in migraine following pattern adaptation extend to simultaneous tilt illusion. Spat Vis 2001; 16: 33–43.
Stanley J. Computer simulation of a model of habituation. Nature 1976; 261: 146–8.
Weiller C, May A, Limmroth V, Jüptner M, Kaube H, von Schayck R, et al. Brain stem activation in spontaneous human migraine attacks. Nature Med 1995; 1: 658–60.
Wray SH, Mijovic-Prelec D, Kosslyn SM. Visual processing in migraineurs. Brain 1995; 118: 25–35.
Young WB, Oshinsky ML, Shechter AL, Gebeline-Myers C, Bradley KC, Wassermann EM. Consecutive transcranial magnetic stimulation: phosphene thresholds in migraineurs and controls. Headache 2004; 44: 131–5.(Gianluca Coppola, Michel Vandenheede, La)
2 Neuroanatomy, University of Liege, CHR Citadelle, Liege, Belgium
3 Headache Clinic, INM Neuromed, Pozzilli (IS), Italy
Summary
A deficit of habituation in cortical information processing, including somatosensory evoked potentials (SSEPs), is the most consistent neurophysiological abnormality in migraine patients between attacks. To explore further the mechanisms underlying this interictal neural dysfunction, we have studied the high-frequency oscillations (HFOs) embedded in SSEPs because they are thought to reflect spike activity in thalamo-cortical cholinergic fibres (early HFOs) and in cortical inhibitory GABAergic interneurons (late HFOs). Untreated migraine patients with (MA) and without (MO) aura were recorded during (n = 13: nine MO, four MA) and between attacks (n = 29: 14 MO, 15 MA) and compared with healthy volunteers. SSEPs were filtered off-line (digital band-pass between 450 and 750 Hz) to extract the two HFO bursts from the broad-band contralateral N20 somatosensory cortical response obtained by median nerve stimulation. In both migraine groups, amplitudes and latencies of conventional broad-band SSEPs recorded interictally from cervical and parietal active electrodes were not significantly different from those found in healthy volunteers. In contrast, maximum peak-to-peak amplitude and area under the rectified curve of the early HFO burst were significantly smaller in both MA and MO patients than in healthy volunteers. There was no significant difference in the later HFO burst between migraineurs and healthy volunteers. During attacks, all electrophysiological measurements in migraineurs were similar to those found in healthy volunteers. Thalamo-cortical activation, as reflected by the early SSEP HFO burst, may thus be reduced in migraine interictally, but normalizes during an attack, whereas intracortical inhibition, as indexed by the late HFO burst, is normal at any time. This supports the hypothesis that the habituation deficit in migraineurs is due to a reduced pre-activation level of sensory cortices and not to increased cortical excitability or reduced intracortical inhibition.
Key Words: high-frequency oscillations; somatosensory evoked potentials; habituation; migraine; cortical excitability
Abbreviations: HFO = high-frequency oscillation; MA = migraine with aura; MO = migraine without aura; SSEP = somatosensory evoked potential
Introduction
The best reproducible interictal electrophysiological abnormality in migraine patients with (MA) or without aura (MO) is lack of habituation (or even potentiation) of evoked and event-related cortical potentials during repetition of the same stimuli (see review by Schoenen et al., 2003). Such a habituation deficit was also reported for somatosensory evoked potentials (SSEPs; Ozkul and Uckardes, 2002). It is most probably responsible for the strong intensity dependence of auditory evoked cortical responses (IDAP) found in migraineurs between attacks (Ambrosini et al., 2003) and for the increased amplitude of evoked potentials reported in some studies (Gawel et al., 1983; Schoenen et al., 2003). Moreover, it has a familial character (Sandor et al., 2002).
The underlying cause of the habituation deficit in cortical information processing is still under debate. In theory, it could be due to increased cortical excitability (and homosynaptic facilitation) or to decreased intracortical inhibition (heterosynaptic inhibition) (Stanley, 1976), both of which have been proposed in migraine following transcranial magnetic stimulation (Aurora et al., 1998; Aggugia et al., 1999; Mulleners et al., 2001; Battelli et al., 2002; Young et al., 2004) and psychophysical studies of the visual cortex (Chronicle and Mulleners, 1994; Wray et al., 1995). Conversely, both methods have also provided results favouring hypoexcitability of the visual cortex and normal function of inhibitory cortical interneurons (afra et al., 1998; Shepherd, 2001; Shepherd et al., 2001; Bohotin et al., 2002, 2003). We have therefore hypothesized, in line with the ‘ceiling theory’ (Knott and Irwin, 1973), that an interictal reduction of the pre-activation level of sensory cortices could be the culprit in migraine. It could indeed explain the low amplitudes found in the first blocks of evoked responses and the lack of habituation in subsequent blocks since it would offer a larger range for suprathreshold cortical activation up to the ‘ceiling’ where habituation starts (Schoenen, 1996). The pre-activation excitability level of sensory cortices is set by activity in thalamo-cortical loops and aminergic, in particular serotoninergic and noradrenergic, projections from the upper brainstem (Mesulam, 1990; Hegerl and Juckel, 1993).
Using digital high-pass filtering (>400 Hz), it was shown in recent years that high frequency (±600 Hz) oscillations (HFOs) of low amplitude are superimposed on the N20 component of median nerve SSEPs. Two bursts of HFOs have been identified in animals (Ikeda et al., 2002; Baker et al., 2003), and with depth (Klostermann et al., 1998, 1999) as well as scalp recordings in humans (Gobbele et al., 1998, 1999, 2003; Hashimoto et al., 1999; Halboni et al., 2000; Haueisen et al., 2001). Early HFOs are thought to be generated by thalamo-cortical afferents, and late HFOs by inhibitory interneurons in parietal area 3b. A study of SSEP HFOs in migraine may thus shed some light on the neurophysiological causes of the interictal habituation deficit found in evoked cortical responses, as it explores simultaneously thalamo-cortical activation and intrinsic inhibition of the parietal cortex. We reasoned that on the one hand the early burst of HFOs would be increased if migraineurs have cortical hyperexcitability between attacks, but decreased if thalamo-cortical pre-activation is reduced. On the other hand, defective intracortical inhibition would result in a decrease of late HFOs.
Material and methods
Subjects
We recorded 42 migraine patients: 19 MA [International Classification of Headache Disorders, II (ICHD-II), Headache Classification Committee of the International Headache Soiety (2004) code 1.2; mean age: 31 ± 11 years.; 11 women, eight men] and 23 MO (ICHD-II code 1.1; mean age: 30.0 ± 10; 17 women and six men). They were compared with 15 healthy volunteers of comparable age and gender distribution (healthy volunteers; mean age 29 ± 7; eight women and seven men) without personal or familial history of migraine and devoid of any detectable medical condition. Subjects taking any medication on a regular basis except for the contraceptive pill were excluded.
We analysed separately the recordings from two groups of patients. In the first (interictal) group (n = 29; 15 MA, 14 MO, 17 females, 12 males), we included only data from patients who had an interval of at least 3 days between the recording and their last or their next migraine attack. The second group (n = 13; 4 MA, 9 MO, 11 females, two males) was composed of subjects in whom the recordings were made within a time range of 12 h before or after the beginning of an attack.
All participants gave informed consent in accordance with the Declaration of Helsinki and the study was approved by the Ethics Committee of the Faculty of Medicine, University of Liege.
Data acquisition
SSEPs were elicited after electrical stimulation of the right median nerve at the wrist using a constant current square wave pulse (0.2 ms width, cathode proximal) and a stimulus intensity set at twice the motor threshold. One series of 500 sweeps was collected and averaged at a repetition rate of 4.4 Hz. The active electrodes were placed over the contralateral parietal area (C3', 2 cm posterior to C3 in the international 10–20 system) and on the fifth cervical spinous process (Cv5), both referenced to Fz; the ground electrode was on the right arm. The evoked potential signals were amplified by CEDTM 1902 pre-amplifiers and recorded by a CEDTM 1401 device (Cambridge Electronic Design Ltd, Cambridge, UK). Subjects were seated relaxed on a comfortable chair in an illuminated room and asked to stay with their eyes opened and to fix their attention on the wrist movement. Forty milliseconds of the post-stimulus period were sampled at 5000 Hz to collect a wide range of frequencies. All recordings were averaged off-line using the SignalTM software package version 2.14 (CED Ltd).
Data analysis
Wide-band somatosensory evoked potentials
For wide-band unfiltered SSEPs (0–2500 Hz), we have identified the various components according to their respective latencies. Thereafter we have measured peak-to-peak amplitudes of the N13 cervival component under the Cv5 active derivation and of N20onset–N20peak, N20peak–P25, P25–N33 under the active C3' left-sided scalp electrode.
High-frequency oscillations
Digital zero-phase shift band-pass filtering between 450 and 750 Hz (Barlett-Hanning window, 51 filter coefficients) was applied off-line in order to extract the HFOs superimposed on the N20 left parietal component of the SSEP. In the majority of recorded traces, we were able to identify two separate bursts of HFOs: the first, early burst occurred in the latency interval of the ascending slope of the conventional N20 component, and the second, late burst in the time interval of the descending slope of N20, sometimes extending into the ascending slope of the N33 peak. In general, the frequency of oscillations was higher in the first than in the second HFO burst, and in between the first and the second burst there was a clear frequency and amplitude decrease, which allowed separation of the two bursts. In recordings in which a clear distinction between the two components was not possible, we considered as first burst HFOs occurring before the N20 peak and as second burst those after the N20 peak.
After elimination of the stimulus artefact, we have measured the latency of the negative oscillatory maximum, the intra-burst frequency, the number of negative peaks, the maximum peak-to-peak amplitude and the area under the rectified curve (AUC), computed using an interval of ±1.17 ms around the maximum peak to provide stability against noise (see Gobbele et al., 1998). All measurements were performed separately on the two HFO bursts.
Finally, since the recorded signal was short (50 ms) and contained high-frequency oscillations that varied with time, we have used a short-time Fourier transform spectrum (STFT) to obtain an indication of the time–frequency properties. This method offers the advantage of visualizing the power (z-axes) of a two-dimensional signal represented by time (x-axes) and frequency (y-axes).
The difference between each group of migraine patients and healthy volunteers was evaluated by one-way analysis of variance (ANOVA), and a P < 0.05 was considered significant.
Results
All the measured components, for both wide-band and HFO SSEPs, were clearly identified in all examined subjects. Tables 1 and 2 show synoptically mean values ± 1 SD and levels of significance found for the various measures taken in healthy volunteers, in interictal recordings from MA and MO patients and in ictal recordings.
Interictal recordings
Broad-band SSEPs
In the MO and the MA groups, peak-to-peak amplitudes and latencies recorded from the cervical and parietal active electrodes were not significantly (P > 0.05) different from those of healthy volunteers.
High-frequency oscillations in SSEPs
Maximum peak-to-peak amplitude and area under the rectified curve of the first burst of HFOs were significantly smaller in MO [F(1,27) = 4.822, P = 0.037 and F(1,27) = 5.456, P = 0.027, respectively] and MA patients [F(1,28) = 7.358, P = 0.011 and F(1,28) = 7.067, P = 0.013] than in healthy volunteers.
In contrast, there was no significant difference between healthy volunteers and patients in amplitude or AUC of the second HFO burst.
Latency of the negative oscillatory maximum, intra-burst frequency and number of negative peaks for both early and late HFO bursts were similar between the three subject groups.
Ictal recordings
In the 13 migraineurs recorded during an attack or in the immediate peri-attack phase, measurements performed on both broad-band and HFOs SSEPs were not significantly different from those of our healthy volunteer group.
Discussion
In accordance with previous studies (Firenze et al., 1988; de Tommaso et al., 1997), we found no significant difference between migraineurs recorded outside of an attack and healthy volunteers for broad-band conventional SSEPs. This suggests that the somatosensory signal is normally conducted along the dorsal columns up to the contralateral somatosensory cortex. In contrast, after band-pass filtering, the first early burst of high-frequency oscillations embedded in the N20 component of the SSEP was significantly reduced in both MA and MO subjects compared with healthy volunteers, while the second late HFO burst was normal.
To the best of our knowledge, our study is the first to apply the rather recent methodology of SSEP HFOs to research on brain physiology in migraine patients. The results have to be interpreted in the light of available, still incomplete and partially contradictory knowledge on the neural generators involved in HFOs.
As mentioned, the late burst of HFOs is thought to have a postsynaptic cortical origin, probably reflecting the activity of intracortical GABAergic inhibitory interneurons (Hashimoto et al., 1996; Mochizuki et al., 2003). Presynaptic subcortical activity, most probably repetitive action potentials along the terminal segment of thalamocortical fibres, is on the contrary responsible for the early burst of HFOs (Gobbele et al., 1998, 2003; Ikeda et al., 2002; Mochizuki et al., 2003). Multichannel scalp recordings and dipole source analysis in normal subjects have shown that both the thalamic and the cortical components of the early presynaptic component of SSEP HFOs were reduced in amplitude during non-rapid eye movement (REM) sleep with partial recovery during REM sleep, while the later HFO bursts were unchanged or rather enhanced (Halboni et al., 2000). Furthermore, the amplitude of the thalamic component and of the cortical tangential source activity of HFOs was significantly greater with eyes open than with eyes closed (Gobbele et al., 2000). After administration of the central acetylcholinesterase inhibitor rivastigmine, Restuccia et al. (2003) found an increase in amplitude of the cortical HFO component in the 18–28 ms latency range and concluded that cortical HFOs are generated mainly by specialized pyramidal neurons. Considering that cholinergic systems play a pivotal role in sleep–wake cycle and arousal (Mesulam, 1995), cholinergic reticulothalamic pathways could be crucial for the generation of SSEP HFOs.
Our finding of a reduced activity of the early HFO burst in migraineurs between attacks suggests, therefore, that migraine is associated with hypofunction in thalamo-cortical excitatory cholinergic afferents. By the same token, the normal second HFO burst indicates that the late generators function normally, be they cortical inhibitory interneurons (Hashimoto et al., 1996; Mochizuki et al., 2003) or pyramidal chattering cells (Restuccia et al., 2003).
As mentioned in the Introduction, controversy has arisen about the excitability status of the cerebral cortex in migraineurs between attacks. Intuitively, some clinical features of migraine such as hypersensitivity to light or noise, and the aura symptoms suggest that the cerebral cortex is hyperexcitable. In the physiological sense, however, hyperexcitability would mean that the brain responds to a subliminal stimulus or that its response to a supraliminal stimulus is increased in amplitude. These two abnormalities have never been demonstrated precisely in migraineurs. Increased amplitudes of psycho-physical (Wray et al., 1995) or cortical evoked responses (Gawel et al., 1983) were reported by some, but these studies are based on the use of repetitive stimuli and we and others have shown that the lack of habituation of an initial low amplitude response during stimulus repetition is responsible for the net increase in averaged responses (Schoenen et al., 2003). Unfortunately, transcranial magnetic stimulation of the visual cortex in migraineurs has provided contradictory results (see the review by Schoenen et al., 2003).
We have therefore hypothesized that the finding of reduced habituation of evoked responses in migraine is probably not due to hyperexcitability and homosynaptic facilitation as such, nor to lack of heterosynaptic inhibition, but to a reduced pre-activation excitability level of sensory cortices (Schoenen, 1996). According to the ‘ceiling theory’ (Knott and Irvine, 1973), this can explain the low response amplitude for a single or a few stimuli and its marked increase (potentiation) or lack of decrease, i.e. habituation, because it allows a wider range of suprathreshold activation before reaching the ‘ceiling’ and initiating habituation of the response. Our finding of reduced early HFO bursts in SSEPs, and thus reduced thalamo-cortical activation is in agreement with this hypothesis and can hardly be explained by increased cortical excitability. Similarly, the normal late SSEP HFO suggests that intracortical inhibition is normal in migraine.
Interestingly, cortical evoked potentials habituate normally just before and during a migraine attack (afra et al., 2000) which suggests, according to the previous ‘ceiling’ hypothesis, that cortical excitability increases to normal levels. As expected, the early HFO burst was within normal values in migraineurs recorded during an attack, which probably reflects an increase in thalamo-cortical activation. Whether such an increase could be induced by the ictal activation of the dorsal rostral brainstem (Weiller et al., 1995; Bahra et al., 2001) which contains state-setting aminergic nuclei projecting to thalamus and cortex remains to be determined. Activation of these nuclei and of thalamo-cortical projections may contribute to an ictal normalization of the cortical pre-activation level.
Supplementary studies of HFO are needed and of interest in migraine. Comparative recordings in other primary headaches such as tension-type headache would be useful to verify the specificity of our findings for migraine. Studying the effect of pharmacological agents, among which prophylactic anti-migraine drugs and repetitive transcranial magnetic stimulations which are able to modify in opposite directions the excitability of the underlying cortex depending on stimulation frequency, can be used to modulate HFOs and shed light on the precise neuronal generators.
References
afra J, Mascia A, Gerard P, Maertens de Noordhout A, Schoenen J. Interictal cortical excitability in migraine: a study using transcranial magnetic stimulation of motor and visual cortices. Ann Neurol 1998; 44: 209–15.
afra J, Sandor PS, Schoenen J. Habituation of visual and intensity dependence of auditory evoked cortical potentials tend to normalize just before and during the migraine attack. Cephalalgia 2000; 20: 714–9.
Aggugia M, Zibeti M, Febbraro A, Mutani R. Transcranial magnetic stimulation in migraine with aura: further evidence of occipital cortex hyperexcitability. Cephalalgia 1999; 19: 465.
Ambrosini A, Rossi P, De Pasqua V, Pierelli F, Schoenen J. Lack of habituation causes high intensity dependence of auditory evoked cortical potentials in migraine. Brain 2003; 126: 2009–15.
Aurora SK, Ahmad BK, Welch KMA, Bdardhwaj P, Ramadan NM. Transcranial magnetic stimulation confirms hyperexcitability of occipital cortex in migraine. Neurology 1998; 50: 1111–4.
Bahra A, Matharu MS, Buchel C, Frackowiak RS, Goadsby PJ. Brainstem activation specific to migraine headache. Lancet 2001; 357: 1016–7.
Baker SN, Curio G, Lemon RN. EEG oscillations at 600 Hz are macroscopic markers for cortical spike bursts. J Physiol 2003; 550: 529–34.
Battelli L, Black KR, Wray SH. Transcranial magnetic stimulation of visual area V5 in migraine. Neurology 2002; 58: 1066–9.
Bohotin V, Fumal A, Vandenheede M, Bohotin C, Schoenen J. Excitability of visual V1–V2 and motor cortices to single transcranial magnetic stimuli in migraine: a reappraisal using a figure-of-eight coil. Cephalalgia 2003; 23: 264–70.
Bohotin V, Fumal A, Vandenheede M, Gerard P, Bohotin C, Maertens de Noordhout A, et al. Effects of repetitive transcranial magnetic stimulation on visual evoked potentials in migraine. Brain 2002; 125: 912–22.
Chronicle E, Mulleners W. Might migraine damage the brain Cephalalgia 1994; 14: 415–8.
de Tommaso M, Sciruicchio V, Tota P, Megna M, Guido M, Genco S, et al. Somatosensory evoked potentials in migraine. Funct Neurol 1997; 12: 77–82.
Firenze C, Del Gatto F, Mazzotta G, Gallai V. Somatosensory-evoked potential study in headache patients. Cephalalgia 1988; 8: 157–62.
Gawel M, Connolly JF, Clifford Rose FC. Migraine patients exhibit abnormalities in the visual evoked potential. Headache 1983; 23: 49–52.
Gobbele R, Buchner H, Curio G. High-frequency (600 Hz) SEP activities originating in the subcortical and cortical human somatosensory system. EEG Clin Neurophysiol 1998; 108: 182–9.
Gobbele R, Buchner H, Scherg M, Curio G. Stability of high-frequency (600 Hz) components in human somatosensory evoked potentials under variation of stimulus rate—evidence for a thalamic origin. Clin Neurophysiol 1999; 10: 1659–63.
Gobbele R, Waberski TD, Kuelkens S, Sturm W, Curio G, Buchner H. Thalamic and cortical high-frequency (600 Hz) somatosensory-evoked potential (SEP) components are modulated by slight arousal changes in awake subjects. Exp Brain Res 2000; 133: 506–13.
Gobbele R, Waberski TD, Thyerlei D, Thissen M, Darvas F, Klostermann F, et al. Functional dissociation of a subcortical and cortical component of high-frequency oscillations in human somatosensory evoked potentials by motor interference. Neuroscience 2003; 350: 97–100.
Halboni P, Kaminski R, Gobbele R, Züchner S, Waberski TD, Herrmann CS, et al. Sleep stage dependent changes of the high-frequency part of the somatosensory evoked potentials at the thalamus and cortex. Clin Neurophysiol 2000; 111: 2277–84.
Hashimoto I, Mashiko T, Imada T. Somatic evoked high-frequency magnetic oscillations reflect activity of inhibitory interneurons in the human somatosensory cortex. EEG Clin Neurophysiol 1996; 100: 189–203.
Hashimoto I, Kimura T, Fukushima T, Iguchi Y, Saito Y, Terasaki O, et al. Reciprocal modulation of somatosensory evoked N20 m primary response and high-frequency oscillations by interference stimulation. Clin Neurophysiol 1999; 110: 1445–51.
Haueisen J, Schack B, Meier T, Curio G, Okada Y. Multiplicity in the high-frequency signals during the short-latency somatosensory evoked cortical activity in humans. Clin Neurophysiol 2001; 112: 1316–25.
Headache Classification Committee of the International Headache Society. International classification of headache disorders, II. Cephalalgia 2004; 24 (Suppl 1).
Hegerl U, Juckel G. Intensity dependence of auditory evoked potentials as an indicator of central serotonergic neurotransmission: a new hypothesis. Biol Psychiatry 1993; 33: 173–87.
Ikeda H, Leyba L, Bartolo A, Wang Y, Okada YC. Synchronized spikes of thalamocortical axonal terminals and cortical neurons are detectable outside the pig brain with MEG. J Neurophysiol 2002; 87: 626–30.
Klostermann F, Nolte G, Losch F, Curio G. Differential recruitment of high frequency wavelets (600 Hz) and primary cortical response (N20) in human median nerve somatosensory evoked potentials. Neurosci Lett 1998; 256: 101–4.
Klostermann F, Funk T, Vesper J, Curio G. Spatiotemporal characteristics of human intrathalamic high-frequency (>400 Hz) SEP components. Neuroreport 1999; 10: 3627–31.
Knott JR, Irwin DA. Anxiety, stress and the contingent negative variation. Arch Gen Psychiatry 1973; 29: 538–41.
Mesulam MM. Large-scale neurocognitive networks and distributed processing for attention, language and memory. Ann Neurol 1990; 28: 597–613.
Mesulam MM. Cholinergic pathways and the ascending reticular activating system of the human brain. Ann NY Acad Sci 1995; 9 (Suppl 2): 23–8.
Mochizuchi H, Machii K, Terao Y, Furubayashi T, Hanajima R, Enomoto H, et al. Recovery function of and effects of hyperventilation on somatosensory evoked high-frequency oscillation in Parkinson's disease and myoclonus epilepsy. Neurosci Res 2003; 46: 485–92.
Mulleners WM, Chronicle EP, Palmer JE, Koehler PJ, Vredeveld JW. Visual cortex excitability in migraine with and without aura. Headache 2001; 41: 565–72.
Ozkul Y, Uckardes A. Median nerve somatosensory evoked potentials in migraine. Eur J Neurol 2002; 9: 227–32.
Restuccia D, Della Marca G, Valeriani M, Rubino M, Paciello N, Vollono C, et al. Influence of cholinergic circuitries in generation of high-frequency somatosensory evoked potentials. Clin Neurophysiol 2003; 114: 1538–48.
Sandor P, Ambrosini A, Agosti RM, Schoenen J. Genetics of migraine: possible links to neurophysiological abnormalities. Headache 2002; 42: 365–77.
Schoenen J. Deficient habituation of evoked cortical potentials in migraine: a link between brain biology, behavior and trigeminovascular activation Biomed Pharmacother 1996; 50: 71–8.
Schoenen J, Ambrosini A, Sandor PS, Maertens de Noordhout A. Evoked potentials and transcranial magnetic stimulation in migraine: published data and viewpoint on their pathophysiologic significance. Clin Neurophysiol 2003; 114: 955–72.
Shepherd AJ. Increased visual after-effects following pattern adaptation in migraine: a lack of intracortical excitation Brain 2001; 124: 2310–8.
Shepherd AJ, Palmer JE, Davis G. Increased visual after-effects in migraine following pattern adaptation extend to simultaneous tilt illusion. Spat Vis 2001; 16: 33–43.
Stanley J. Computer simulation of a model of habituation. Nature 1976; 261: 146–8.
Weiller C, May A, Limmroth V, Jüptner M, Kaube H, von Schayck R, et al. Brain stem activation in spontaneous human migraine attacks. Nature Med 1995; 1: 658–60.
Wray SH, Mijovic-Prelec D, Kosslyn SM. Visual processing in migraineurs. Brain 1995; 118: 25–35.
Young WB, Oshinsky ML, Shechter AL, Gebeline-Myers C, Bradley KC, Wassermann EM. Consecutive transcranial magnetic stimulation: phosphene thresholds in migraineurs and controls. Headache 2004; 44: 131–5.(Gianluca Coppola, Michel Vandenheede, La)