Abstract
Amplitude and habituation of event-related potentials are abnormal in migraine. We investigated 43 migraine and 41 healthy families to evaluate the influences of age, sex and familial contribution on the variance of amplitude and habituation of the contingent negative variation (CNV). Analysis of individual differences in relation to the CNV habituation was performed. The study demonstrated that habituation of the early CNV component characterizes migraine considerably better than the CNV amplitudes. Habituation, however, is strongly influenced by age. Migraine adults and children generally showed reduced habituation. Surprisingly, more than 30% of the healthy adults demonstrated a marked loss of habituation. The reduced CNV habituation represented a high sensitivity but low specificity to migraine, especially in children. CNV amplitude and habituation parameters revealed a considerable familial contribution associated with migraine. No familial influence on either morphology or habituation of the CNV in healthy families or between healthy members of migraine families was observed. The low specificity and familial transmission of CNV parameters in members of migraine families suggest that increased amplitudes and reduced habituation of CNV do not constitute a primary risk factor for migraine, but rather represent a predisposition. Genetic components probably affect variation of the CNV amplitude and habituation.
Keywords
Introduction
Among the basic pathophysiological elements of migraine, much attention has been directed to the responsiveness to sensory stimuli. Various studies have shown that migraine patients are characterized by increased amplitudes and reduced habituation of event-related potentials (ERPs). These abnormalities were demonstrated in a large number of conditions: in migraine with and without aura, in children and adults, using visual and acoustic stimulation, and in slow cortical potentials as well as early and late complexes of evoked potentials (1–11). Large ERP amplitudes and reduced habituation seem to represent a satisfactory specificity to migraine since they were found by comparison of migraineurs not only with healthy controls but also with tension-type headache patients, having normal or low amplitudes and normal habituation (3, 8, 9, 10, 12). These abnormalities are closely related to migraine pathogenesis since they are more pronounced a few days before a migraine attack, representing a neurophysiological disposition of the brain to the headache (5, 7). Moreover, they normalize following effective treatment with prophylactic anti-migraine medication (13–15). Therefore, increased amplitudes and reduced habituation of ERPs imply a generalized dysfunction in cortical information processing in migraine, which is often discussed as a specific feature of the migrainous brain (1, 3, 16).
Investigations of amplitude and habituation of ERPs in migraine have not answered a number of important questions. For example, the sensitivity and specificity of these abnormalities in this type of headache have rarely been quantitatively evaluated. While reduced habituation of ERPs is discussed as a method with high potential diagnostic value (3), its prevalence in the population of migraine patients was insufficiently investigated. Evers et al. (4) and Lahat et al. (8) reported high specificity of abnormalities in amplitude and habituation of visual ERPs in migraine children compared with age-matched healthy controls and tension-type headache sufferers. The diagnostic value of acoustically evoked potentials and slow brain potentials, and the specificity of the described processing abnormalities in adults are, however, unknown. This information is of special importance since there is no sufficient diagnostic physiological marker of migraine available for difficult cases (16).
It has been shown that the clinical course of migraine and the responsiveness to acute and prophylactic anti-migraine treatment in children differ from adult migraineurs (17, 18). However, little is known about physiological mechanisms of juvenile migraine. Brain maturation may possibly influence migraine-specific neurophysiological abnormalities in childhood (4, 8). There are very few studies that have taken the processes of maturation into consideration. On the one hand, it was demonstrated that the neurophysiological abnormalities of migraine children are age-dependent and resemble those of adult migraineurs (4, 10). On the other hand, it is well known that children generally are characterized by large, especially negative, amplitudes of ERPs and by reduced habituation compared with adults (19, 20). Interaction between features of sensory maturation and migraine-specific physiological features of information processing must be investigated in more detail because of the potential significance for migraine pathogenesis.
Genetic influences were described for both EEG and ERPs (21, 22). It has been suggested that complex diseases, such as migraine, are not appropriate starting points for molecular genetic research (23). Partial factors ‘closer to gene action’, such as EEG traits and features of ERPs, seem to be fitting (22). Functional-genetic studies are very promising and their effectiveness has already been proved (24, 25). There are indications that increased amplitudes and reduced habituation of ERPs in migraine are genetically determined and represent an appropriate target for functional-genetic studies. A preliminary study performed in our laboratory demonstrated that not only children suffering from migraine, but also their healthy siblings show increased amplitudes of slow cortical potentials (26). Pronounced cortical negativity is neither necessary nor sufficient alone for the occurrence of migraine, but rather represents a predisposition. The results of this investigation suggest familial transmission of increased cortical negativity. Sandor et al. (27) studied related pairs of parents and children suffering from migraine and analysed the degree of similarity of evoked potentials using the Monte Carlo method. They found that parents showed recordings significantly more similar to those of their own children than to children of other parents. This was true both for the amplitude/stimulus intensity function slope of auditory evoked cortical potentials and the habituation of pattern-reversal visually evoked potentials, supporting an underlying familial factor in cortical information processing. However, both studies had the following limitations: they did not analyse both the amplitudes and habituation, there was no control group, and variables such as age and sex were not considered.
The present study set out to answer the following questions: how sensitive and specific are alterations of slow cortical potentials in migraine children and adults? What influences do age and sex have on contingent negative variation (CNV) abnormalities? and what is the familial contribution to the variance of CNV features in migraine and healthy families?
In order to answer these questions the CNV was investigated in migraine and healthy families. The CNV is a slow cortical potential which can be recorded from the scalp between two stimuli while the subject is anticipating the second event and preparing for task performance, i.e. dependent on contingency (28). It has been demonstrated that the CNV is associated with almost all neuropsychologic processes with resource mobilization such as attention, expectation, motivation, emotion and others (28, 29). From a physiological point of view, the CNV amplitude reflects the excitability of the dentritic trees of cortical pyramidal neurones following activation in the thalamo-cortical loop (29). More than a decade of research has emphasized that CNV is a suitable method for studying information processing abnormalities in migraine (2, 6, 7, 9, 10, 12, 14, 15, 26).
Subjects and methods
Subjects
Forty-three families with one child suffering from migraine without aura and 41 healthy families took part in the study. Table 1 shows demographic and clinical characteristics of the participants. Children suffering from migraine were recruited from the out-patient Department of the Clinic of Neuropaediatrics (University of Kiel). Patients with the clear diagnosis ‘Migraine with or without aura’, without co-morbidity or abnormal epileptiform EEG data were selected from the data collected in the Clinic of Neuropaediatrics during the last 7 years. The chosen families (n = 60) were informed about the study by letter and contacted several weeks later for information about the actual course of the migraine and co-morbidity. Only children with at least one migraine attack without aura per month during the last 3 months, and without chronic tension-type headache, were included in the study (n = 48). Headache was diagnosed in accordance with the criteria of the International Headache Society (IHS; (30), code 1.1.). Because of the poor reliability of the IHS criteria for paediatric migraine (18), care was taken that the migraineurs were characterized by (i) vomiting during migraine attacks, (ii) a history of migraine in the family, and (iii) severe headaches with pronounced disability subjectively evaluated on the visual-analogue scale (from 0 (neither) to 10 (extremely severe)) between 8 and 10. The children were first diagnosed by a paediatric neurologist from the Clinic of Neuropaediatrics and then by a neurologist of the Department of Medical Psychology. Parents suffering from migraine were diagnosed by neurologists from the Department of Medical Psychology. The adults suffered from migraine without aura. Since recruitment was based on archive data, all of the migraineurs had a migraine history longer than 2 years. Forty-three of the 48 investigated families were included in the statistical analysis. The five families were excluded because of poor data quality, extremely frequent eye movement artefacts, or lack of compliance of family members. The migraineurs were asked to keep a headache diary during the month prior to recording. This allowed assessment of the influence of clinical parameters on the CNV and determination of the investigation day considering the periodicity phenomenon (5, 7). Care was taken that the time interval between the attacks and the recordings was at least 5 days. The interval between recording and the next attack was registered at the follow-up visit or by telephone contact. The recording was repeated if an attack occurred within 5 days; this was the case in 16 subjects.
Demographic and clinical characteristics of the groups
According to demographic parameters the groups differ significantly with respect to the number of men (χ2, P < 0.05 for all comparisons), which is substantially lower in the group of adult migraineurs and healthy siblings of migraine children, and the number of pubertal and post-pubertal children (χ2, P < 0.05 for all comparisons), which differ significantly in the group of healthy siblings from the distribution in both other groups of children. According to the age no differences between groups of children and between groups of adults were observed (
Healthy children and their parents were recruited from two German schools. The families were checked for neurological and psychiatric disorders not only in the participants, but also in the two past generations. The families did not suffer from migraine, neurologic or psychiatric disorders, drug or alcohol abuse in any of the analysed generations and had no major medical problems. The healthy families were matched for age and sex as far as possible. There were no significant differences with regard to age (Table 1, one-way
The participants had not used prophylactic migraine medication for at least 6 months prior to the investigation and did not have any neurologic, psychiatric or other disorders (including acute infections). The acute migraine medication was only considered to exclude the drug-induced headache. The influence of the actual state of the participants on the CNV was considered by recording the quality and duration of sleep during the night before investigation, mood (subjective evaluation on the digital scale from 0 to 9), medication, drug, alcohol and caffeine intake on the day of recording using a standardized questionnaire. There were no differences between the groups with regard to any of these factors (one-way
The study was permitted by the Ethic Committee of the Faculty of Medicine, University of Kiel, Germany. The subjects were informed about the course of the experiment and gave written informed consent according to the Helsinki convention.
Experimental design and recordings
All participants were seated in an armchair with open eyes in an electrically shielded sound-attenuated room with dimmed light. The auditory warning (S1) and imperative (S2) stimuli with an intensity of 75 dB(a) were produced by a loudspeaker located behind the subject. The interval between S1 and S2 was 3 s. A CNV session consisted of 32 trials in which the subject had to react to the imperative stimulus (GO-response). Additionally, eight trials were randomly presented where no reaction was expected (NO-GO-response). The warning stimulus (S1) for the GO-response had a frequency f = 1000 Hz and lasted 100 ms. The warning tone for the NO-GO-response had a frequency f = 200 Hz. The imperative stimulus (S2) had a frequency f = 2500 Hz, lasted a maximum of 1500 ms and was deactivated by pressing a button. Reaction time was defined as the period between onset of S2 and the pressing of a button. S1 and S2 pairs were offered at random intervals of 10–15 s. The duration of one recording was 6 s (the recording began 1 s before S1 and ended 2 s after S2). The period between recording onset and S1 was taken as the baseline for all measurements.
The EEG was recorded using non-polarizable Ag/AgCl electrodes over Cz according to the International 10–20 System with linked mastoids as reference. The electrode site on the scalp was prepared with alcohol and scraped with rough paper, resulting in an electrode impedance of < 7 kΩ. The EEG signals were amplified using a Nihon Kohden amplifier with a time constant of 5 s (equivalent to the low frequency filter of 0.03 Hz) and high frequency filter of 35 Hz and digitized at a rate of 100 Hz for each channel. Vertical eye movement artefacts were excluded by parallel recording of the electrooculogram (EOG) using electrodes (Ag/AgCl) positioned 1–1.5 cm above and below the right eye. The trial was rejected if EOG deflections > 20 μV interfered with 5 s of the EEG recording. A protocol listed the number of rejected trials for each recording. Approximately 8% of the EEG trials were rejected due to eye movement artefacts. There were no significant differences between groups in the number of rejected trials.
Data analysis
All CNV recordings were analysed by two trained raters blind to the subject group. The interrater reliability was r = 0.94. The blinded evaluation was performed since some of the CNV recordings had to be analysed off-line. The trials with movement or EMG artefacts, not registered by the EOG trigger, were rejected (2.1 + 1.4 trials in each recording were removed off-line, especially in the groups of children). The GO-trials were averaged and the amplitudes of the total CNV, early and late components and post-imperative negative variation (PINV) were calculated. The total CNV was assessed between 500 ms and 3000 ms following S1. The early CNV component was defined as the mean amplitude in a window of 200 ms duration around the maximal amplitude of the expectancy wave between 550 and 750 ms after S1 (2). The late component was the mean amplitude during the 200 ms preceding S2. PINV was the mean amplitude of CNV between 500 and 2000 ms following S2.
Each recording was divided into eight blocks of four consecutive trials to determine the course of habituation and trends in the CNV amplitudes. Habituation was indicated by a negative, whereas dishabituation was marked by a positive slope calculated by linear regression (y = ax + b, where a is the slope of habituation and b the intercept of linear regression) (6, 7, 31). This study considered only the early CNV component habituation data. This limitation of data presentation is based on the literature (6, 7) and on observations in this sample of patients, whereby only the early CNV showed habituation differences between groups (
Statistical evaluation
Since the data were normally distributed (non-significant results in the Kolmogorov–Smirnov test) and characterized by homogenous variances (F-test), the analysis of variance was performed with either the General Lineal Model (GLM) or the one-way
Results
CNV amplitudes
Table 2 shows mean amplitudes of different components of the CNV. Only the amplitude of the early CNV component differed significantly between the groups (one-way
Mean amplitudes (μV) of the total contingent negative variation (CNV), the early, the late CNV components and the post-imperative negative variation (PINV), reaction time and descriptive characteristics of habituation presented as a coefficient of lineal regression ‘a’ (y = ax + b) and initial value ‘b’ in members of migraine and healthy families
Post hoc Scheffé tests:
P < 0.05
P < 0.01 for comparisons with both groups of healthy adults. No other comparisons were significant.
Figures in parentheses–standard deviation.
CNV habituation
Figure 1 shows the course of habituation (the change in the amplitude of the early CNV component over eight trial blocks) within the groups. Table 2 demonstrates the mean coefficient of linear regression for the eight CNV blocks (slope of habituation, ‘a’). Analysis of variance with repeated measure (CNV amplitude in eight recording blocks) and the between-groups factor ‘group of subjects’ demonstrated that the groups differed significantly according to the course of the CNV habituation (F (35, 1430) = 1.74, P = 0.005). Only healthy adults were characterized by habituation of the early CNV component (negative amplitude/block of recording function slope in Table 2). The most pronounced dishabituation was observed in the group of migraine children. However, both groups of children and migraine adults demonstrated positive CNV slopes. In order to determine whether migraine or age had a greater influence on the abnormal CNV habituation, an analysis of variance with the within-groups factor habituation (CNV amplitude in eight blocks of recording) and the between-groups factors ‘age’ (children vs. adults), ‘disease’ (migraine vs. healthy) and ‘sex’ (male vs. female) was performed. Age (F (7, 2002) = 3.44, P = 0.001) and presence of disease (F (7, 2002) = 2.72, P = 0.008) showed main effects. The sex of the subject did not exert any influence on the CNV habituation (F (7, 2002) = 0.75, P = 0.626).
Amplitude of the early contingent negative variation (CNV) component (μV) in eight trial blocks of CNV recording and slope of CNV habituation described as a linear regression in family members of migraine and healthy families.
Analysis of the differences between the groups according to the slope of the CNV habituation confirmed the observation that migraine patients and children generally are characterized by a pronounced loss of CNV habituation. One-way
Discriminant analysis performed with the independent variables ‘age’, ‘diagnosis’, ‘sex’ and amplitude of the early CNV component demonstrated predictive values of these variables for habituation (patients with habituation vs. patients with dishabituation). The χ2 test for the first discriminate function was significant (χ2 = 44.042, P < 0.001). The first canonical correlation was 0.4. The most predictive variable (canonical coefficient, canonical loading) for (loss of) habituation was ‘diagnosis’ –(0.59, 0.73). Age of subjects and amplitude of the early CNV component demonstrated lower predictive values (0.45, 0.62 and 0.43, 0.60, respectively). Sex did not have any predictive value. Migraine exerts a more pronounced influence on habituation of the early CNV component than age and may explain a larger part of the variation of CNV habituation.
Analysis of individual differences
Differentiation between subjects with habituation and those with dishabituation was achieved using a cut-off based on the regression coefficient (‘a’ in y = ax + b), characterizing the slope of habituation as an amplitude/block of recording function. A person was classified as showing habituation if a < 0. Dishabituation was assumed if a ≥ 0 (8). Figure 2 shows the absolute and relative numbers of subjects demonstrating habituation or dishabituation for each group of subjects. In accordance with the previous results, patients suffering from migraine and children generally are normally characterized by dishabituation. Both groups of children and the group of adult migraineurs differed significantly from the healthy adults (χ2 > 4.32, d.f. = 1, P < 0.04 for all comparisons).
Prevalence of individuals which could be classified as showing contingent negative variation (CNV) habituation (coefficient of linear regression a < 0) or as demonstrating loss of CNV habituation (coefficient of linear regression a ≥ 0) in each group of members of migraine and healthy families.
Analysis of individual differences enabled evaluation of sensitivity and specificity of dishabituation in relation to migraine. Of migraine children 75.6%, and of adult migraineurs 80.0% were characterized by abnormal habituation of the CNV. However, more than 60% of healthy children and approximately 35% of the healthy adults (from migraine and healthy families) showed the same abnormal pattern of information processing. Therefore, reduced early CNV component habituation shows a satisfactory sensitivity but poor specificity to migraine.
Familial contribution
A correlation analysis was performed in order to evaluate similarity among family members for CNV morphology and habituation. Based on the ERP studies of twins and families, it was expected that family members would demonstrate more similar waveform morphology than unrelated individuals (32, 33). To prove this, 84 ‘non-families’ of unrelated subjects were constructed from the real families by random distribution of the members. This randomized distribution was performed separately for migraine (n = 43) and healthy (n = 41) families. The members of such a ‘non-family’ were not related biologically each other. These ‘non-families’ served as controls in the correlation analyses (32). The Pearson's product-moment correlations were performed between pairs of family members in migraine and healthy families separately, as well as in ‘non-families’. Figure 3a, b demonstrates the correlation coefficients between related family members of migraine families and healthy families for the early CNV component. The only significant correlations were observed for the amplitude (Fig. 3a) and habituation of the early CNV between migraine children and parents suffering from migraine (r = 0.42; P < 0.05 for CNV habituation), as well as between migraine children and healthy parents with familial history of migraine (first-degree relatives of these parents suffering from migraine) (r = 0.6; P < 0.05 for CNV habituation). Similarities in CNV amplitude were more pronounced (strong positive correlations) than in CNV habituation (moderate positive correlations). No significant correlations between migraine children and healthy parents without familial migraine, or between members of healthy families and ‘non-families’ according to amplitude and habituation of the early CNV were found (Fig. 3b). Small but significant correlations of the late CNV component were observed between children and parents in healthy families (r = 0.34; P = 0.026) and healthy members of migraine families (r = 0.23; P < 0.034).
Pearson's product-moment correlations (scatter plots) for the amplitude of the early contingent negative variation (CNV) component between children suffering from migraine, their parents suffering from migraine and healthy parents with a positive family history of migraine (a), and children suffering from migraine and their healthy parents without a positive family history of migraine as well as between related subject from healthy families (b). Correlation coefficients r are given. No significant correlations between healthy family members from migraine families and subjects from non-families were observed. Correlations for the habituation of the early CNV component are explained in the text.
The significance of similarities in amplitude and habituation of the early CNV component between migraine children and their parents and of the amplitude of the late CNV component between related healthy subjects was examined. The correlation coefficients obtained from the comparisons—father vs. migraine child, mother vs. migraine child, father vs. healthy sibling, mother vs. healthy sibling, migraine child vs. healthy sibling in migraine families, and father vs. child, and mother vs. child in healthy families—were subjected to a Fisher z transformation to normalize the distribution (as described in (34)). An analysis of variance (General Lineal Model with Greenhouse–Geisser corrections (∊) for sphericity violations) with the within-subjects factor ‘genetic relation’ (real family vs. ‘non-family’) and between-subjects factor ‘type of family’ (migraine vs. healthy) was performed separately for the early and late CNV components and for the slope of habituation from the transformed correlation values (taken as cases). We only observed a significant effect of ‘genetic relation’ (F (1, 2) = 46.987, P = 0.002) and significant interaction ‘genetic relation × type of family’ (F (1, 2) = 8.185, P = 0.039) for the amplitude of the early CNV component and a tendency towards significance for the interaction ‘genetic relation × type of family’ (F (1, 2) = 6.223, P = 0.059) for the slope of habituation. Effects or interactions of the late CNV component were non-significant. The most pronounced similarities between CNV parameters were related to the amplitude of the early CNV component. These similarities were more pronounced between migraine children and parents also suffering from migraine, or having a familial history of migraine. The slope of habituation demonstrated similarities but these were not so marked as for the amplitude of the early CNV.
Influencing variables
In order to investigate the influence of age and clinical course of migraine on CNV amplitudes and habituation, correlation analyses were performed. No significant correlations were found between CNV characteristics and the clinical parameters presented in Table 1. However, there was a small but significant negative correlation between age and slope of habituation (Pearson's product-moment correlation coefficient, r =−0.36, P < 0.001) in healthy subjects. This means the greater the age, the more pronounced is the CNV habituation. The correlation between age and the CNV habituation slope in migraine patients was less pronounced but also significant (r =−0.23, P < 0.046). Surprisingly, the age-related correlation of the amplitude of the early CNV differed between migraine and healthy subjects. Negative significant correlations between age and CNV amplitude could be described (r =−0.26, P = 0.021) for migraineurs. Not only the steepest positive slope of habituation, but also the largest CNV amplitudes were found in migraine children, especially in the youngest of them. Correlations between age and the early CNV component were weak and non-significant in healthy subjects (r = 0.13, P = 0.052). No correlation between age and the late CNV component was found.
Discussion
Habituation of the early CNV demonstrated a greater pathogenetic significance for migraine than its amplitude or other CNV components. Patients did not differ from age-matched healthy controls with respect to CNV amplitudes. However, migraine was closely associated with reduced habituation of the early CNV. Schoenen et al. (35) and Kropp & Gerber (6) demonstrated that increased CNV amplitudes are explained by reduced habituation. This was confirmed by other authors, showing reduced habituation as an important abnormality of information processing in migraineurs (1, 3–5, 11). Even if migraine undergoes transformation into chronic daily headache, the underlying pathophysiologic changes manifest in a reduction of CNV amplitudes, but not in normalization of the reduced CNV habituation (36). Some prophylactic agents also affect the CNV amplitudes, but not habituation (15). This neurophysiological abnormality seems to represent the most stable and characteristic feature of information processing in migraine (16).
Habituation of the early CNV component was characterized by high sensitivity and low specificity to migrainous headache in the present study: 75.6% of the migraine children and 80% of adult migraineurs demonstrated a pronounced reduction of the CNV habituation. This is in accordance with Evers et al. (4), who found reduced habituation of the P300 latency and amplitude in more than 70% of migraine children. However, 60% of the healthy children and 35% of the healthy adults also showed impaired CNV habituation in our study. This lowers the diagnostic value of reduced CNV habituation and hinders the use of this method in clinical practice. The prevalence of reduced habituation in the healthy population, with significant association of this physiological trait to migraine (the most predictive value for loss of habituation in the discriminant analysis was presence of migraine in a particular subject), suggests predisposition of these people to migrainous headache. Reduced habituation of CNV does not constitute a primary risk factor for transmission of the disorder, but is rather a contributing risk factor.
Apart from migraine, age appears to be the most important variable predicting individual differences and extent of CNV habituation. The majority of the children were characterized by reduced habituation of the early CNV. Significant correlations between age and habituation and age and CNV amplitude (early component) were found in both migraine and healthy families. This seemed to be due to changes in younger years, since there was no significant age effect for the amplitude and CNV habituation slope within the age range 14–55 years. The habituation of the early CNV component is more pronounced and the CNV amplitude less negative with increasing age, representing processes of sensory maturation. This conforms with the observation that children generally show more pronounced negativity, manifesting in large amplitudes of slow cortical potentials and long- and short-latency event-related potentials (20, 37). It can be suggested that puberty is a critical period resulting in the acceleration of sensory maturation, so that children and adolescents older than 14–16 years do not differ from adults according to any CNV parameters (34). Surprisingly, very few studies of sensory maturation have analysed habituation of ERPs. Together with the studies of Freedman et al. (19) and Evers et al. (4), our investigation emphasizes the key role of reduced habituation in brain maturation, individual development and structuralization of information processing. Since normal values of information processing differ according to age, this must always be taken into consideration in studies of ERP habituation.
Analysis of age-related distribution of CNV amplitude and habituation provides indirect support for the association of migraine with cortical excitability (38). The amplitudes of negative potentials decrease and positive potentials increase with increasing age (20, 37). ERP negativity is interpreted as cortical excitability, facilitating the processing of sensory input, while ERP positivity is a manifestation of neuronal inhibition (28, 29). Sensory maturation results in predominance of inhibitory over facilitatory processing, in increasing inhibition of cortical excitation (37). Since cortical excitability and the reduction of CNV habituation in children are more pronounced than in adults, it can be proposed that the reduced CNV habituation is associated with increased cortical excitability. Moreover, if normalization of CNV habituation with increasing age does not occur in migraine patients this may be related to the increased cortical excitability in migraineurs. Biochemical studies of the activity of excitatory amino acids, MEG investigations, the identification of low magnesium levels by means of MR spectroscopy, and transcranial magnetic stimulation confirm the assumption that the migrainous brain is hyperexcitable (for review see (38)), although there are some opposite findings (39). However, there are no direct investigations of the relationship between ERP habituation and cortical excitability. Further research is needed to clarify and integrate this relation in the puzzle of migraine pathogenesis.
In the present study correlation analysis showed a possible familial contribution to the variance of the CNV amplitude and habituation. It demonstrated that there are close similarities in morphology and habituation of the early CNV component between migraine children and parents with migraine. Moreover, we found similarities between migraine children and healthy parents with first-degree relatives suffering from migraine. No similarities in morphology or habituation were observed between members of healthy families, non-related individuals or healthy relatives in migraine families. These results emphasize that familial factors contribute to a degree to the variance of CNV parameters and that this contribution is migraine-related. The role of migraine history in structuralization of familial CNV similarities suggests that familial transmission of CNV morphology and habituation is possibly associated with migraine-vulnerability genes. This is in accordance with the results of Sandor et al. (27) on the intensity dependence of auditory evoked cortical potentials and the habituation of pattern-reversal visually evoked potentials. The degree of genetic contribution to migraine, however, could not be assessed. Twin studies and investigations of large families with adult migraineurs are needed to evaluate the genetic component of the variation of abnormal CNV amplitudes and habituation and to describe the pattern of familial transmission. The present study emphasized two important points: (i) the family contributes to abnormal information processing in migraine, and increased amplitudes and reduced habituation of ERPs can be used in functional-genetic studies of this form of headache; and (ii) the influence of environmental factors is large and has to be taken into account in migraine.
It is surprising that the present study did not confirm previously shown abnormalities in CNV amplitude (2, 6, 7, 9, 10), but emphasized the reduced CNV habituation as a pathogenetic feature of the migraine brain. The results demonstrate that familial influences associated with migraine are more likely to be found for the CNV amplitude than for the habituation as discussed by Sandor et al. (27). These differences may be in part due to statistical bias. The amplitude of the early CNV component was characterized by larger variance than the CNV habituation slope in our study. It is well known that the larger the variance, the greater the statistical reliability and possibility of strong significant correlations, and the smaller the chance of significant group differences because of the small test power, and vice versa (40). The most reliable test for verification of the results described in this study is cross-validation (replication of results in an independent sample of subjects), which will be undertaken in future.
It can be concluded that reduced habituation of the early CNV component represents a non-specific feature of migrainous information processing. This feature is neither necessary nor alone sufficient for the development of migraine, but constitutes a risk factor. The transmission of CNV characteristics in migraine families may be genetically determined. However, additional primary genetic and/or environmental factors contribute to a large degree to the occurrence of migraine. These additional factors are especially important in children, since reduced CNV habituation is characteristic of normal information processing in childhood. Analysis of CNV parameters describes migraine, at least in adults, as a delayed sensory maturation possibly genetically influenced and related to impaired cortical excitability.
Footnotes
Acknowledgement
This study was supported by the German Research Foundation, Grant Ge 500/4-1, 4-2.