On the Pathophysiology of Migraine Links for Empirically Based Treatment With Neurofeedback

Transcription

1 Applied Psychophysiology and Biofeedback, Vol. 27, No. 3, September 2002 ( C 2002) On the Pathophysiology of Migraine Links for Empirically Based Treatment With Neurofeedback Peter Kropp, 1,2 Michael Siniatchkin, 1 and Wolf-Dieter Gerber 1 Psychophysiological data support the concept that migraine is the result of cortical hypersensitivity, hyperactivity, and a lack of habituation. There is evidence that this is a brain-stem related information processing dysfunction. This cortical activity reflects a periodicity between 2 migraine attacks and it may be due to endogenous or exogenous factors. In the few days preceding the next attack slow cortical potentials are highest and habituation delay experimentally recorded during contingent negative variation is at a maximum. These striking features of slow cortical potentials are predictors of the next attack. The pronounced negativity can be fed back to the patient. The data support the hypothesis that a change in amplitudes of slow cortical potentials is caused by altered habituation during the recording session. This kind of neurofeedback can be characterized as empirically based because it improves habituation and it proves to be clinically efficient. KEY WORDS: migraine; contingent negative variation; neurofeedback. THE PATHOGENESIS OF MIGRAINE ATTACKS Migraine is characterized by intense recurrent, unilateral pain attacks, with a frequency of one or more attacks per month and lasting 4 72 hr. The severity of the pulsating pain and accompanying symptoms (such as nausea, vomiting, photophobia, or phonophobia) can cause a migraineur to cease all activities. During a migraine attack paroxysmal vascular and electrical changes, caused by centrally triggered brainstem disturbances, are observed. Current studies show that this disease is driven by enhanced neuronal activity in the brainstem (Wang & Schoenen, 1998; Weiller et al., 1995; Welch, Cao, Aurora, Wigins, & Vikinstad, 1998). The relation between neuronal and vascular processes is emphasized by the theory of neurogenic inflammation (Goadsby, 1997). In the current view triggering factors seem to initiate a cascade of central processes, which lead to a migraine attack. The trigeminovascular system plays an important role in the pathogenesis of migraine and in pain processing (Moskowitz, 1993). This abnormal pain processing is often associated with a 1 Institute of Medical Psychology, University of Kiel, Germany. 2 Address all correspondence to Peter Kropp, Institute of Medical Psychology, University of Kiel, Niemannsweg 147, D Kiel, Germany; kropp@med-psych.uni-kiel.de /02/ /0 C 2002 Plenum Publishing Corporation

2 204 Kropp, Siniatchkin, and Gerber cortical excitatory state that enables trigeminovascular nerve fibers to depolarize (Hardebo, 1992). Taken together, the vast majority of the current studies implies a lot of biobehavioral triggering factors among migraine patients. Therefore it is of interest to investigate neurophysiological or biopsychological correlates of this biobehavioral vulnerability. The often described stimulus sensitivity can be investigated by recording evoked potentials (EP) or slow cortical potentials (SCP). PSYCHOPHYSIOLOGY OF MIGRAINE There are two psychophysiological models standing for migraine: First, a biopsychological correlate is the low electro-cortical preactivation level (Schoenen, 1996). This condition of cortical structures causes a large range of supra-threshold activation, which explains the strong intensity dependence of responses with an augmenting pattern and a reduced habituation of the response. A low cortical preactivation level indicates low serotoninergic activity and a deficit in mitochondrial metabolism in migraine (Lodi et al., 1997). Second, is the presence of an increased preactivation level in combination with a higher excitability (Welch & Ramadan, 1995). This theory is supported by observations of increased concentration of excitatory amino acids in migraine, and large amplitudes of ERPs (Aurora, Ahmad, Welch, Bhardhwaj, & Ramadan, 1998). Both models emphasize the abnormality of the cortical preactivation level and the excitability of dendritic trees of cortical pyramidal neurones and they state an impaired threshold regulation in migraine patients (Schoenen, 1998). One method to assess preactivation level and cortical excitability is the recording of the contingent negative variation (CNV). The CNV is a slow event-related potential that is associated with the orienting response and it reflects the allocation of processing resources to mental activities such as expectancy, attention, and preparation of behavior (Maertens de Noordhout, Timsit-Berthier, Timsit, & Schoenen, 1986; Rockstroh, Elbert, Canavan, Lutzenberger, & Birbaumer, 1998). As described by Rohrbaugh et al. (1986) these preparatory processes include stimulus anticipation, response preparation, and decision anticipation. CNV can be observed between two defined stimuli: a warning stimulus (S1) that announces an imperative stimulus (S2) requiring a reaction time response. Between S1 and S2, the overall CNV (tcnv) occurs as a negative DC-shift. For statistical analysis single-trial CNVs are grand-averaged. With regard to migraine pathogenesis, it is interesting that the CNV is believed to be modulated by the central monoaminergic systems (Rohrbaugh et al., 1986). The CNV can be described into three components: (1) the total CNV (tcnv) as the mean amplitude between S1 and S2, (2) the early component of the CNV (amplitude between 550 and 750 ms after S1, icnv), which is used as an index level of expectancy, mediated by the noradrenergic system, and (3) the late component (last 200 ms before S2, lcnv) which reflects motor preparation and mobilization, corresponding to the dopaminergic system (Nagel-Leiby, Welch, D Andrea, Grundfeld, & Brown, 1990). A higher noradrenergic activity induced by stimulation of the locus coeruleus leads to an improvement of signal detection and enhancement of attentional functions. CNV studies in migraine research have revealed several particularities which are discussed below: (a) pronounced negativity and reduced habituation, (b) normalization during the attack, (c) periodicity between the attacks, (d) altered state of cerebral maturation, and (e) voluntary control of amplitudes and habituation by biofeedback.

3 Psychophysiology of Migraine 205 Fig. 1. Grand averaged CNV curves for migraine patients (in the pain free interval, thick line) and of healthy controls (thin line). All records were obtained from Cz with linked mastoids (Kropp & Gerber, 1995). Pronounced Negativity and Reduced Habituation Migraine patients appear to produce increased negative amplitudes of cortical eventrelated potentials compared with healthy controls during the pain-free interval (see Fig. 1). Table I shows CNV amplitudes from a sample of migraine patients (without aura) and sex and age matched healthy controls. The differences in icnv are significant. The stronger negativity in migraine patients CNV indicates a more pronounced intensity of general attention allocation and enhanced mental resources allocated to a task. During the recording session habituation occurs in healthy subjects, leading to a reduced overall mean amplitude, whereas in migraine patients habituation does not occur (Kropp & Gerber, 1993, 1995). The significant increase of icnv amplitude in migraine patients is caused by a reduced habituation. This deficit in habituation kinetics is in line with the lack of habituation found for other evoked potentials (e.g. Evers, Quibelday, Grotemeyer, Suhr, & Husstedt, 1999). Habituation, the simplest type of learning, helps the organism to ignore meaningless stimuli and protects the cortex against over-stimulation. The abnormal habituation supports the concept of migraine being an orienting or an attention dysfunction. It is well known that high arousal is accompanied by disturbed habituation. The finding that migraine patients Table I. Averaged CNV Amplitudes (M)inµV of 32 Trials for the Migraine Group in the Pain-Free Interval, and for the Healthy Control Group a Migraine (n = 16) Healthy (n = 22) M SD M SD Z p tcnv icnv lcnv Note. Statistics were performed with Mann-Whitney U-tests. SD: Standard deviation; tcnv: total CNV; icnv: early CNV-component; lcnv: late CNV-component. a During the pain-free interval data were taken more than 3 days after the last or before the next migraine attack.

4 206 Kropp, Siniatchkin, and Gerber show a reduced but not a complete lack of CNV habituation may suggest a state of high cortical arousal rather than a defect of the mesolimbic DA system as discussed by others (Gerber & Schoenen, 1998). On the other hand, low central serotoninergic transmission could be responsible for the occurrence of response potentiation and the lack of habituation during repetition of sensory stimuli. It seems likely that both inherited and acquired changes in central monoaminergic systems are responsible for the observed cortical hyperreactivity. Additionally, the dramatic failure of brain homeostasis can also be observed by the occurrence of cortical spreading depression during the migraine attack (Lauritzen, 2001). Normalization During the Attack During the attack, migraine patients show no differences in CNV amplitudes or habituation compared with healthy subjects, CNV amplitudes normalize, and icnv habituation does not differ from healthy controls (Kropp & Gerber, 1995, see Fig. 2). Göbel, Krapat, Fig. 2. Grand averaged CNV curves for the migraine group in the pain-free interval, during an attack, and the healthy control group. All records were made from Cz with linked mastoids.

5 Psychophysiology of Migraine 207 Ensink, and Soyka (1993) did not confirm these results, but they used a different CNV methodology. The normalization of CNV during the attack could be explained by two mechanisms: (1) it may be a primary phenomenon due to a homeostatic protective mechanism involving a breakdown of the noradrenergic arousal system, which reduces the threshold for a migraine attack (Welch, 1989), or (2) it could be a reduced attention ability due to the headache interval. According to some authors, the migraine attack may be interpreted as a neurohumoral reaction aimed at protecting the brain from noxious influences. The biobehavioral model by Welch (1989) postulates that strong activators (e.g., stress, specific triggers, mood) may induce potential shifts in the frontal cortex leading to activation of the locus coeruleus by means of an orbitofrontal brainstem pathway. Welch emphasized that chronic stress may lead to a depletion of central noradrenergic storage, whereas avoidable or weak stress and relaxation increases the excretion of NA catabolites. On the other hand, serotoninergic neurons in the raphé nuclei are minimally influenced by stress; moreover, during migraine attacks, increased levels of serotonin have been found in the blood (Ferrari, 1992). Thus, the normalized CNV during the migraine attack could be due to the depletion of noradrenergic activity in combination with increased serotoninergic transmission. We assume that normalization of the CNV during an attack reflects a protective cerebral mechanism and homeostasis secondary to cerebral exhaustion (Kropp & Gerber, 1995), triggered by stress or other individual factors. Periodicity Between the Attacks In the last few days before the next attack, CNV in migraine patients becomes more negative and habituation diminishes, indicating the occurrence of the next attack (Kropp & Gerber, 1998, see Fig. 3(a) and (b). The differences between CNV directly before and the first day following a migraine attack are statistically significant, with a tendency for a more pronounced negativity in CNV before the migraine attack in relation to the mean CNV. In the first day after an attack CNV amplitude is significantly lower than the mean CNV. High CNV amplitudes in the days before an attack possibly indicate an effort to control hyperactive levels of brain intrinsic activity. There is some evidence that this hyperactivity refers to a paroxysmal brain dysfunction causing the migraine attack. The day after the attack amplitudes are lower, indicating a less hyperactive level, possibly caused by a refractory period following the attack. The data suggest that a migraine attack is possible every 3 days, but not more frequently. Therefore a migraine attack may occur when the high negative amplitude of the icnv coincides with other precipitating factors. In some cases high CNV amplitudes can indicate an approaching migraine attack. This periodicity, observable by frequent scp-recordings, has been confirmed by others (Evers et al., 1999). During the days near a migraine attack CNV amplitudes in migraine patients will vary over a wide range. This has to be taken into consideration in future CNV research because CNV data have to be interpreted with regard to the distance of the next attack. Altered State of Cerebral Maturation Examination of CNV across various age ranges and populations provides some insights into migraine pathogenesis. The CNV was used to record the influence of age in

6 208 Kropp, Siniatchkin, and Gerber Fig. 3. (a) Grand averaged CNV amplitudes in µv related to the 5 days before (day 5 today 1) and 5 days after (day +1 today+5) a migraine attack (ma); n: number of observations, a: reference line indicating mean CNV amplitude of µv. (b) CNV amplitudes in µv of 32 trials for the migraine group grand averaged in the 5 days before a migraine attack (pre) and 5 days after an attack (post), recorded on the day before a migraine attack ( 1) and the day after a migraine attack (+1), recorded on the last 2 days before ( 2/ 1) and after the attack (+1/+2), compared between all records except during an attack (all) and 1 day before the attack ( 1) respectively 1 day after the attack (+1). 162 migraine patients and 320 healthy controls aged between 8 and 59 years. Six age groups were created, and the mean values of icnv were analysed (Kropp, Siniatchkin, Stephani, & Gerber, 1999). Up to the age group of years, migraine patients and healthy controls revealed similar high icnv amplitudes and habituation slopes. In the age

7 Psychophysiology of Migraine 209 Fig. 4. Grand averaged icnv amplitudes in V for 162 migraine patients and 320 healthy controls in six age groups between 8 and 59 years ( p < 0.01, p < 0.001). group of years, migraine patients and older healthy controls showed pronounced habituation and a decline in icnv amplitudes (Fig. 4). Adult migraine patients revealed the same pronounced cortical excitability observed both in normal and migraine children. These higher CNV amplitudes indicate that adult migraine patients show stronger attentional efforts, without habituation, over the trials within a CNV session. This deficiency in habituation corresponds to the state necessary in an early ontogenetic state, where increased facilitation of information processing is needed. Regarding the results of the CNV analysis, the observed assumption of a maturation effect in icnv parameters suggests two developmental stages. The first stage, up to the age group of years, is characterized by high icnv both in healthy controls and in migraine patients. The second stage is observed only in healthy adults. This stage can be explained as development of the basis for habituation abilities and selective information processing (Oades, Dittmann-Balcar, & Zerbin, 1997). The second stage begins with a jump in amplitude and habituation of icnv. There are similarities with observations in EEG complexity, which show maximum gain during puberty as an indication of development of the frontal cortex and its cortico cortical connections (Anokhin, Birbaumer, Lutzenberger, Nikolaev, & Vogel, 1996). The results suggest an enhanced attentional effort in adult migraine patients, observed by a absence of habituation.

8 210 Kropp, Siniatchkin, and Gerber Synopsis The above presented psychophysiological aspects of migraine show that there are observable abnormalities in the pain-free interval in migraine patients characterized by a pronounced CNV negativity and the absence of habituation. The migraine attack seems to be the end of this phasic disturbance of cortical or subcortical function during the last days before the attack. The low cortical preactivation level and the high excitability seem to be key-factors for the understanding of the pathogenesis of migraine. Direct influence on these factors can lead to clinical improvement of migraine. Prophylactic antimigraine medication diminishes the possibility for the occurrence of the next attack by influencing cortical excitability and preactivation level. For example, Schoenen (1986) and Ahmed (1999) recorded diminished CNV amplitudes in migraine patients after beta-blocker treatment. This normalization in CNV amplitudes corresponded with clinical improvement of the frequency and duration of migraine attacks. Can the same be accomplished by self-regulation strategies? INTEGRATION: FROM PSYCHOPHYSIOLOGY TO EMPIRICALLY BASED NEUROFEEDBACK The mechanism of self-regulation and instrumental conditioning of slow cortical potentials can be explained by the regulation of thresholds of cortical excitability and preactivation (Elbert, 1993). Voluntary increase of cortical negativity causes better performance in arithmetic tasks, faster reaction times, larger amplitudes of acoustic startle-reflex, and better discrimination of visual stimuli (Brody et al., 1994; Lutzenberger, Roberts, & Birbaumer, 1993). Impaired self-regulation of activity of neuronal networks may be caused by neurological or psychiatric disorders (Rockstroh et al., 1993). The aim of empirically-based neurofeedback is the attempt to control SCPs directly and voluntarily by thoughts, cognitions, and imagery to achieve a normalization of CNV amplitudes. This voluntary control of SCPs should lead to adequate stimulus processing and a normalized habituation function in CNV recordings. In a recently published study (Siniatchkin et al., 2000), 10 children suffering from migraine without aura learned to control SCPs within 10 feedback sessions. All sessions followed an identical format. The children received 20 trials of baseline CNV recording (reaction time paradigm, interstimulus-interval (ISI) t = 3 s). After a break, increase of SCP negativity within the ISI was required for the next 30 trials. The last 30 trials involved suppression of SCP negativity. The change in feedback-strategies (increase and decrease) was introduced to enhance a wide variety of amplitude effects. After this biofeedback procedure, transfer trials without visual feedback (15 increase of negativity, 15 decrease of negativity) were arranged. The feedback signal appeared on a video monitor 5 s after S2. The amplitude of the SCP (mean amplitude in the time window between 500 and 3000 ms, recording from the vertex with linked mastoids as reference) was presented as a blue stripe moving to the right or to the left according to the change in brain activity. The length of the stripe was a linear function of the integrated EEG referenced to the mean of the 2 s pretrial baseline. The criterion for success of self-regulation for each session was defined as the level of success achieved during the previous session. During the first two sessions, the migraine children compared with healthy controls were characterized by a lack of ability to control cortical negativity, especially during transfer trials. However, there was no difference in the

9 Psychophysiology of Migraine 211 Fig. 5. Changes in the clinical course of migraine in relation to the pretreatment data in the treatment group and the waiting list group (100%: data obtained before the biofeedback therapy or observation period; frequency: number of migraine attacks; hatched line: waiting list group). subsequent training sessions. Overall, feedback training was accompanied by a significant reduction of cortical excitability. Both healthy children and children suffering from migraine were able to influence the amplitude of SCPs. However, migraine children demonstrated good learning abilities with regard to SCP self-regulation, especially during feedback trials, and they did not differ from healthy children after 10 training sessions. Migraine children were better able to suppress their cortical negativity, in relation to the actual baseline, than to increase it. During the training, they learned to increase SCPs more effectively. Feedback training was accompanied by a significant reduction of the actual baseline from session to session and it was clinically effective (Fig. 5). The SCP training led to a significant reduction of days with migraine and other headache parameters compared to the state revealed before the training and compared to the clinical course of migraine in a wait list group. The efficacy of self-regulation of SCPs emphasizes the importance of slow cortical activity in the pathogenesis of migraine because reduction of the SCP amplitude correlates with a marked clinical improvement. We assume that direct control of the CNV amplitudes via self-regulation may have a preventive effect. Possibly SCP-feedback improves the internal control of cortical preactivation and excitability. The effect could be a reduction of the high CNV amplitudes. Effective strategies to control SCPs correlated with a marked improvement of habituation during the CNV recording. Therefore we assume that the voluntary reduction in negative amplitudes is the effect of better habituation abilities. Further studies should focus on the development of strategies to enhance this basic skill. SUMMARY Revealed by SCPs, migraine can be regarded as a disorder related to a habituation deficit caused by a brainstem-related dysfunction. Habituation shows systematic variations around a migraine attack. After an attack, habituation is normal. With neurofeedback, migraine patients can learn to control their high negative SCPs and to habituate. Along with the ability to habituate the days with migraine can be reduced.

10 212 Kropp, Siniatchkin, and Gerber REFERENCES Ahmed, I. (1999). Contingent negative variation in migraine: Effect of beta blocker therapy. Clinical Encephalography, 30, Anokhin, A. P., Birbaumer, N., Lutzenberger, W., Nikolaev, A., & Vogel, F. (1996). Age increases brain complexity. Electroencephalography and Clinical Neurophysiology, 99, Aurora, S. K., Ahmad, B. K., Welch, K. M. A., Bhardhwaj, P., & Ramadan, N. M. (1998). Transkranial magnetic stimulation confirms hyperexcitability of occipital cortex in migraine. Neurology, 50, Brody, S., Rau, H., Köhler, F., Schupp, H., Lutzenberger, W., & Birbaumer, N. (1994). Slow cortical potential biofeedback and the startle reflex. Biofeedback and Self-Regulation, 19, Elbert, T. (1993). Slow cortical potentials reflect the regulation of cortical excitability. In W. C. McCallum & H. Curry (Eds.), Slow potentials in the human brain (pp ). New York: Plenum. Evers, S., Quibelday, F., Grotemeyer, K.-H., Suhr, B., & Husstedt, I.-W. (1999). Dynamic changes of cognitive habituation and serotonin metabolism during the migraine interval. Cephalalgia, 19, Ferrari, M. D. (1992). Biochemistry of migraine. Pathology Biology, 40, Gerber, W.-D., & Schoenen, J. (1998). Biobehavioral correlates in migraine: The role of hypersensitivity and information-processing dysfunction. Cephalalgia, 18, Goadsby, P. J. (1997). Current concepts of the pathophysiology of migraine. Neurology Clinics, 15, Göbel, H., Krapat, S., Ensink, F. B., & Soyka, D. (1993). Comparison of contingent negative variation between migraine interval and migraine attack before and after treatment with sumatriptan. Headache, 33, Hardebo, J. E. (1992). A cortical excitatory wave may cause both the aura and the headache of migraine. Cephalalgia, 12, Kropp, P., & Gerber, W.-D. (1993). Is increased amplitude of contingent negative variation in migraine due to cortical hyperactivity or to reduced habituation? Cephalalgia, 13, Kropp, P., & Gerber, W.-D. (1995). Contingent negative variation during migraine attack and interval: Evidence for normalization of slow cortical potentials during the attack. Cephalalgia, 15, Kropp, P., & Gerber, W.-D. (1998). Prediction of migraine attacks using a slow cortical potential, the contingent negative variation. Neuroscience Letters, 257, Kropp, P., Siniatchkin, M., Stephani, U., & Gerber, W.-D. (1999). Migraine evidence for a disturbance of cerebral maturation in man? Neuroscience Letters, 276, Lauritzen, M. (2001). Cortical spreading depression in migraine. Cephalalgia, 21(7), Lodi, R., Montagna, P., Soriani, S., Iotti, S., Araldi, C., Cortelli, P., et al. (1997). Deficit of brain and sceletal muscle bioenergetics and low brain magnesium in juvenile migraine: An in vivo 31P magnetic resonance spectroscopy study. Pediatric Research, 42, Lutzenberger, W., Roberts, L. E., & Birbaumer, W. (1993). Memory performance and area-specific self-regulation of slow cortical potentials: Dual-task interference. International Journal of Psychophysiology, 15, Maertens de Noordhout, A., Timsit-Berthier, M., Timsit, M., & Schoenen, J. (1986). Contingent negative variation in headache. Annals of Neurology, 19, Moskowitz, M. A. (1993). Neurogenic inflammation in the pathophysiology and treatment of migraine. Neurology, 41, Nagel-Leiby, S., Welch, K. M., D Andrea, G., Grundfeld, S., & Brown, E. (1990). Event-related slow potentials and associated catecholamine function in migraine. Cephalalgia, 10, Oades, B. A., Dittmann-Balcar, A., & Zerbin, D. (1997). Development and topography of auditory event-related potentials (ERPs): Mismatch and processing negativity in individuals 8 22 years of age. Psychophysiology, 34, Rockstroh, B., Elbert, T., Birbaumer, N., Wolf, P., Düchting-Röth, A., Reker, M., et al. (1993). Cortical selfregulation in patients with epilepsies. Epilepsy Research, 14, Rockstroh, B., Elbert, T., Canavan, A., Lutzenberger, W., & Birbaumer, N. (1998). Slow cortical potentials and behavior (2nd ed.). Baltimore: Urban & Schwarzenberg. Rohrbaugh, J. W., McCallum, W. C., Gaillard, A. W., Simons, R. F., Birbaumer, N., & Papakostopoulos, D. (1986). ERPs associated with preparatory and movement-related processes. A review. Electroencephalography and Clinical Neurophysiology, 38, Schoenen, J. (1986). Beta blockers and the central nervous system. Cephalalgia, 6, Schoenen, J. (1996). Abnormal cortical information processing between migraine attacks. In M. Sandler, M. Ferrari, & S. Harnett (Eds.), Migraine: Pharmacology and genetics (pp ). London: Altman. Schoenen, J. (1998). The pathophysiology of migraine: A review based on the literature and on personal contributions. Functional Neurology, 1, Siniatchkin, M., Hierundar, A., Kropp, P., Kuhnert, R., & Gerber, W.-D. (2000). Self-regulation of slow cortical potentials in children with migraine: An exploratory study. Applied Psychophysiology and Biofeedback, 25(1),

11 Psychophysiology of Migraine 213 Wang, W., & Schoenen, J. (1998). Interictal potentiation of passive oddball auditory event-related potentials in migraine. Cephalalgia, 18, Weiller, C., May, A., Limmroth, V., Jupner, M., Kaube, H., Schayck, R. V., et al. (1995). Brain stem activation in spontaneous human migraine attacks. Nature Medicine, 1, Welch, K. M. A. (1989). The search for a universal hypothesis of migraine mechanisms. The Cummings Lecture. In F. Clifford-Rose (Ed.), New advances in headache research. Nishimura: Smith-Gordon. Welch, K. M. A., Cao, Y., Aurora, S., Wigins, G., & Vikinstad, E. M. (1998). MRI of the occipital cortex, red nucleus, and substantia nigra during visual aura of migraine. Neurology, 51, Welch, K. M. A., & Ramadan, N. M. (1995). Mitochondria, magnesium and migraine. Journal of Neurological Sciences, 134, 9 14.