Sage Journals: Discover world-class research

Abstract

Increased negative amplitudes and lack of habituation of contingent negative variation (CNV) in migraine are well established and are supposed to reflect an altered cortical excitability level. Migraine attacks occur less during pregnancy but often relapse after delivery. We investigated the effect of pregnancy on slow cortical potentials and reaction time in migraine patients and healthy controls. Four groups were examined: 14 pregnant migraine patients, 12 non-pregnant migraine patients, 15 pregnant healthy women and 16 non-pregnant healthy women aged 19-38 years. Two recordings were performed in the pregnant subjects: in the 36th week of gestation and 4 weeks after delivery. The nonpregnant subjects were recorded at the same time interval of 8 weeks. Pregnant migraine patients showed significantly fewer migraine days during the third trimester of pregnancy and returned to nearly the former level 4 weeks post delivery. Non-pregnant migraine patients demonstrated a significant reduction of migraine days at the second measurement. There was no effect of pregnancy on CNV amplitudes, but there was an effect of pregnancy on the habituation coefficient and reaction time of migraine patients. Faster habituation from a higher preactivation level was found. As an explanation for the changed habituation level we favour the model of correlation between preactivation level and habituation level, the so-called law of initial value. We found a correlation between preactivation level and habituation. Our study confirms a specific effect of pregnancy on slow cortical potentials in migraine patients.

Keywords

Migraine pregnancy contingent negative variation

Introduction

Migraine often improves during pregnancy, an observation confirmed by several epidemiological studies. Most of these studies have been made retrospectively (1–8). A few studies investigated patients prospectively (9–12). Diagnostic headache diaries during patients’ pregnancy until 1 month after delivery have been used (12). Migraine attacks have been shown to be reduced during pregnancy, but often reoccurring after delivery (12).

The pathophysiology of migraine has been widely examined with electrophysiological methods. The brain of the migraine patient is characterized by two different electrophysiological models related to interictal sensory information processing. The first model favours a high cortical preactivation level in combination with higher cortical excitability (13). The second model explains a low cortical preactivation level by reduced habituation of the cortical response due to hypofunctioning subcortico-cortical aminergic pathways (14, 15). A low cortical preactivation level would be due to low serotoninergic activity and a changed mitochondrial metabolism in migraine (16). Effects of repetitive transcranial magnetic stimulation on visual evoked potentials in migraine indicate that migraineurs could be characterized by reduced cortical excitability interictally (17).

A method to determine the interictal activation level and cortical excitability is to record the contingent negative variation (CNV) (18). The CNV is a slow event-related potential associated with the orienting response and reflects the allocation of processing resources to mental activities such as expectancy, attention and preparation of behaviour (19, 20).

Many studies have recorded increased amplitudes of the overall CNV (oCNV) as well as the initial CNV (iCNV) after the first stimulus and lack of habituation of iCNV between migraine attacks (21, 22). There is evidence that cortical preactivation levels change with a longer history of migraine (23).

The aim of the present study was to examine the effect of pregnancy on slow cortical potentials in migraineurs. Significant reduction of migraine attacks is often achieved by medical prophylaxis with β-blocking agents. The clinical effects manifest themselves with improved habituation of the iCNV (24). Pregnant migraineurs often reduce their frequency of migraine attacks during pregnancy (12). Therefore, we assume that pregnant migraineurs behave as patients after prophylactic treatment. We hypothesize that pregnant migraine patients tend to normalize iCNV amplitudes and habituation. We expect lower iCNV amplitudes and changed habituation level during pregnancy and a relapse after delivery.

Materials and methods

Patients and participants

Four groups of participants took part in the study: 14 pregnant migraine patients, 12 non-pregnant migraine patients and two control groups: 15 pregnant healthy women and 16 non-pregnant healthy women. One of the pregnant and one of the non-pregnant migraine patients were excluded because of missing data. For the pregnant women and patients two recordings were made: in the 36th week of gestation and 4 weeks after delivery. Non-pregnant women and patients participated in the recordings at the same time interval (8 weeks).

The non-pregnant healthy control participants without any history of primary headache or neuromuscular disease were recruited from students, whereas the pregnant healthy subjects were found through newspaper advertisements. Pregnant migraine patients were recruited from several gynaecological practices and through the newspaper, whereas non-pregnant migraine patients were acquired from a specialized neurological practice in Kiel, Germany. All migraine patients had a migraine diagnosis made by a neurologist according to the revised International Headache Society (IHS) criteria (25). They all filled out diagnostic headache diaries for 8 weeks, and none of them had received prophylactic treatment for at least 3 months. The pregnant migraine patients suffered from 2.9 migraine days per month over the last 12 months, 100% had regularly taken analgetics and 46.7% triptan medication before pregnancy. The non-pregnant migraine patients suffered from 3.2 migraine days per month over the last 12 months, 92.3% took analgesics and 30.8% triptan medication before measurement. According to the IHS criteria (25), the migraine patients had suffered from migraine with aura (MA) or migraine without aura (MoA) over at least 3 years with a minimum of one attack per month in the last 3 months before pregnancy and measurement. The clinical characteristics of the four groups are shown in Table 1. The patients’ symptomatology was classified as ‘improved’ if there was a reduction in the frequency of migraine days by ≥ 50% compared with the situation before pregnancy (26), as ‘worsened’ if the frequency of migraine days showed an increase of > 50%.

Table 1

Clinical characteristics

	Non-pregnant healthy women	Non-pregnant migraine patients	Pregnant healthy women	Pregnant migraine patients
n	16	12	15	14
Type of migraine (MoA/MA)	—	10 MoA/2 MA	—	9 MoA/5 MA
Age (years)	24.7 ± 5.3	30.9 ± 7.9	33.7 ± 4.1	33.1 ± 4.0
Duration of migraine (years)	—	13.4 ± 9.2	—	15.8 ± 6.6
Age at initial manifestation (years)	—	19.0 ± 4.9	—	17.4 ± 5.7
Intensity of attack (1–3)∗	—	2.8 ± 0.4	—	2.9 ± 0.3
Attacks/4 weeks∗	—	1.9 ± 1.5	—	1.9 ± 1.0
Days with migraine/4 weeks∗	—	3.2 ± 2.8	—	2.9 ± 1.5

∗

Migraine characteristics averaged 3 months before pregnancy and measurement.

n, number of patients; MoA, migraine without aura; MA, migraine with aura.

EEG procedure

All EEG recordings were taken in a headache-free interval for at least 72 h before or after an attack. This was secured by an inquiry at the recording date and a telephone call 4 days after the recording. Patients kept headache diaries. A warning stimulus S1 (1000 Hz, 75 dB, duration 100 ms) was presented 3 s before an imperative stimulus S2 (2500 Hz, 75 dB, duration of maximum 1500 ms terminated by pressing a button as soon as possible) was presented. Forty CNV trials were presented as a choice reaction task with 32 GO and eight NO-GO trials. The NO-GO trials (200 Hz, 75 dB, duration 100 ms, no reaction required) were randomly presented to minimize adaptation effects during the recording session (27). NO-GO trials were not recorded or analysed in this study.

Reaction time was measured between S2 and pressing the button. EEG was averaged over 6 s (1 s before S1 until 2 s after S2). We used Ag/AgCl electrodes at Cz according to the International 10–20 System with linked mastoids as reference and with an electrode impedance of < 6 kΩ. The EEG was amplified with a bandpass of 0.03–35 Hz and digitized with a sampling rate of 100 Hz for each channel. The period before S1 was used as a baseline, and eye movements were recorded with an electrooculogram (EOG) using electrodes positioned above and below the right eye (bipolar recording). CNV was rejected if the peak activity of EOG exceeded 50 µV (time constant 0.3 s, frequency < 35 Hz). A protocol listed the number of rejected trials for each subject. In no case did this number exceed n = 5. For data analyses only GO trials were averaged, and reaction time, amplitudes of the oCNV, iCNV and terminal CNV (tCNV) were analysed. NO-GO trials were not analysed.

The iCNV was computed as the peak amplitude in a period of 200 ms within the maximum amplitude between 550 and 750 ms after S1 (28). The mean amplitude between 2800 and 3000 ms after S1 served to calculate tCNV, whereas oCNV was the mean amplitude between S1 and S2.

One block comprised four sequential GO trials, and the entire recording was composed of eight blocks. Habituation of iCNV (calculated with linear regression y = aX + b, where a is a habituation coefficient and b the constant) was also calculated. A positive habituation coefficient arose from decreased amplitudes and represented habituation, whereas a negative habituation coefficient represented increased amplitudes characterizing lack of habituation.

Statistical analysis

Statistical analysis was performed using SPSS (Statistical Package for the Social Sciences, version 10.1; SPSS Inc., Chicago, IL, USA). Initial two-way analyses of variance (anova) and two-tailed post hoc t-tests were conducted to identify the influence of the factors ‘migraine’ and ‘pregnancy’ and to identify between-group differences. We calculated means of iCNV, habituation coefficients and reaction time for each time point. The α-level was corrected according to Bonferroni. For clinical data the α-level was set at 0.05. For CNV parameters the α-level was defined as 0.008.

Results

Clinical data: course of migraine days before pregnancy, during the third trimester and 4 weeks after delivery

In the group of pregnant migraine patients, the number of migraine days per 4 weeks declined significantly (t ₁₃ = 2.6; P = 0.023) during the third trimester of pregnancy, but increased again to nearly the former level within the first 4 weeks post delivery (t ₁₃ = −3.1; P = 0.008). There were no migraine days in 53% (n = 8), an improvement of migraine days in 20% (n = 3), no one with unchanging migraine days and an increase of migraine days in 27% (n = 4) of the cases. Within the 4 weeks after delivery, the attacks reoccurred in 80% (n = 12) of all pregnant migraine patients, in four cases (33%) within the first week after delivery.

In the group of non-pregnant migraine patients, there was also a reduction of migraine days at the first measurement (t ₁₁ = 2.2; P = 0.051), which became significant 8 weeks later at second measurement (t ₁₁ = 3.9; P = 0.002). Table 2 and Fig. 1 show the results.

Figure 1

Frequency of migraine days in the 4 weeks before pregnancy, during the third trimester and 4 weeks post delivery in pregnant and non-pregnant migraine patients.

Table 2

Migraine days per 4 weeks in pregnant and non-pregnant migraine patients

Days with migraine/4 weeks	Pregnant migraine patients			t-value	P	Days with migraine/4 weeks	Non-pregnant migraine patients			t-value	P
Days with migraine/4 weeks	n	M	s.d.	t-value	P	Days with migraine/4 weeks	n	M	s.d.	t-value	P
Before pregnancy	14	2.9	1.5			Prior to recordings	12	3.2	2.8
36 weeks of gestation	14	1.1	1.6	2.5	0.023∗	1. Recording	12	2.1	1.6	2.2	0.051
4 weeks post delivery	14	2.9	3.1	−3.1	0.008∗	2. Recording	12	1.4	1.5	3.9	0.002

n, number of patients; M, mean value of migraine days per 4 weeks; s.d., standard deviation; P, significance level.

Electrophysiological data: effects of pregnancy and migraine on slow cortical potentials

Slow cortical potentials at the first measurement

We found no effect of pregnancy and migraine on oCNV, tCNV or iCNV amplitudes. The anova showed an interaction between pregnancy and migraine (F _1,53 = 6.85, P = 0.012): pregnant migraine patients presented a more pronounced positive habituation coefficient of iCNV amplitudes at the 36th week of gestation than pregnant healthy women (t ₂₇ = −3.36; P = 0.002). There was no effect of pregnancy or migraine on the constant of iCNV at the first measurement. Table 3 shows the results of the CNV components of the four groups examined at the first measurement. Figure 2 presents the linear regressions consisting of the habituation coefficient and constant of iCNV amplitudes of the four groups at first and second measurement.

Figure 2

Linear regressions [habituation coefficients and constants of initial contingent negative variation (iCNV) amplitudes] of the four groups in the 36th week of gestation and 4 weeks after delivery. One block comprises four sequential GO trials, each recording composed of eight blocks.

Table 3

CNV components of the examined four groups at the first measurement

Group		Reaction time (ms)	oCNV (µV)	iCNV (µV)	tCNV (µV)	Habituation coefficient	Intercept
Non-pregnant healthy women	M	294.5	−5.2	−4.3	−8.6	0.4	−6.1
n = 16	s.d.	72.7	4.8	4.7	7.6	1.0	6.4
Non-pregnant migraine patients	M	248.1	−6.4	−7.4	−9.8	−0.3	−6.4
n = 12	s.d.	44.7	4.7	4.3	7.7	0.6	4.9
Pregnant healthy women	M	294.7	−3.9	−5.5	−4.5	−0.4	−5.1
n = 15	s.d.	43.0	3.5	4.3	4.8	1.1	7.3
Pregnant migraine patients	M	350.1	−7.1	−7.7	−8.1	0.3	−8.8
n = 14	s.d.	53.9	6.7	6.3	8.4	1.1	8.7

n, number of patients; M, mean value of migraine days per 4 weeks; s.d., standard deviation; oCNV, overall contingent negative variation; iCNV, initial contingent negative variation; tCNV, terminal contingent negative variation.

There was an interaction effect between pregnancy and migraine on reaction time (F _1,53 = 11.77, P = 0.001). Pregnant migraine patients showed a significantly slower reaction time compared with non-pregnant migraine patients (t ₂₄ = 5.19; P = 0.000) as well as pregnant healthy women (t ₂₇ = 3.09; P = 0.005).

Slow cortical potentials at the second measurement

The second measurement demonstrated an effect of migraine on iCNV amplitudes (F _1,53 = 5.51, P = 0.023). Migraine patients demonstrated more negative iCNV amplitudes than healthy women (t ₂₆ = 3.05; P = 0.005). There was no effect on oCNV or tCNV amplitudes. For the habituation coefficient and the constant of iCNV amplitudes, we saw no effect of pregnancy on migraine at the second measurement. Table 4 shows the results of the CNV components of the four groups at the second measurement.

Table 4

CNV components of the examined four groups at the second measurement

Group		Reaction time (ms)	oCNV (µV)	iCNV (µV)	tCNV (µV)	Habituation coefficient	Intercept
Non-pregnant healthy women	M	255.7	−3.4	−3.4	−7.6	0.0	−3.8
n = 16	s.d.	56.7	5.4	4.4	9.2	0.7	6.2
Non-pregnant migraine patients	M	248.2	−6.9	−9.0	−7.6	0.2	−10.2
n = 12	s.d.	49.3	4.6	5.2	5.7	0.8	5.7
Pregnant healthy women	M	288.1	−6.8	−6.2	−10.0	0.1	−6.4
n = 15	s.d.	49.9	4.3	4.8	6.0	1.4	8.3
Pregnant migraine patients	M	312.0	−6.0	−7.1	−8.4	0.1	−7.6
n = 14	s.d.	59.6	3.0	3.5	4.1	1.2	7.3

We observed an effect of pregnancy on reaction time at the second measurement (F _1,53 = 10.75, P = 0.002). There was a trend to a slower reaction time in migraine patients 4 weeks after delivery compared with migraine patients who where not pregnant (t ₂₄ = 2.86; P = 0.009).

Discussion

Clinical findings

The number of migraine days per 4 weeks dropped significantly in the third trimester of pregnancy and rose to the former level 4 weeks after delivery. These results concur with previous findings (12). This might be the result of lack of hormonal fluctuation throughout pregnancy (3). In contrast, non-pregnant migraine patients also showed a reduction of migraine days at the second measurement. This unexpected effect may be due to effects of improvement of migraine as a result of using diagnostic headache diaries (29). The results suggest an effect of pregnancy on decreasing migraine days, which is probably stronger than the effect of headache diaries.

Electrophysiological findings

The electrophysiological findings show the well-known effect in migraine sufferers of more negative amplitudes of iCNV and a lack of habituation in iCNV amplitudes in non-pregnant migraine patients (22). As expected, migraine patients during the second recording demonstrated more negative iCNV amplitudes and an increased constant.

Furthermore, there are two main findings of our study. The first was an effect of pregnancy on a changed habituation level in migraine patients. Pregnant migraine patients are characterized by habituation compared with pregnant healthy women, who demonstrated a lack of habituation.

What may be the reason of the changed habituation level during pregnancy? There is evidence that the differences in memory, information processing and attention tasks could not be explained by the change of steroid hormones in pregnancy (30, 31). As an explanation for the changed habituation level we favour the model of correlation between preactivation level and habituation level, the so-called law of initial value (32, 33). This theory describes a psychophysiology mechanism of counterregulation dependent upon the preactivation level. The first block of the iCNV represents the cortical preactivation level. We observed a correlation between the first block of iCNV (preactivation level) and habituation coefficient. More negative iCNV amplitudes in the first block led to a more positive habituation coefficient (r = −0.44, P = 0.001). Pregnant migraine patients had high negative iCNV amplitudes in the first block, leading to a positive habituation coefficient. Pregnant healthy women had small negative iCNV amplitudes in the first block and therefore a negative habituation coefficient. This favours on initial value concept. A recent study has shown evidence that cortical preactivation level changes with longer history of migraine. Migraine patients with short-lasting migraine (< 121 months) show little preactivation, whereas long-lasting sufferers (> 120 months) show a higher preactivation level (23). The changed preactivation level seems to reflect long-lasting migraine disease and might result from different initial noradrenergic activity.

Additionally, there is an effect of pregnancy on reaction time, whereas pregnant migraine patients had dramatically slower reaction times than the other groups. Research has demonstrated a decrease in speed of information processing and attention in women in late pregnancy (30). However, in our study there was no effect of reaction time in healthy pregnant women. Only pregnant migraine patients showed slower reaction times. Decelerated reaction times were also found in the Stroop test comparing migraine patients and healthy controls (34). We hypothesize that pregnant migraine patients produce slower reaction times because of their higher preactivation level. According to the above-mentioned concept of initial value, we suppose an effect of preactivation level on reaction performance. A high preactivation level not only affected the habituation coefficient, but seemed also to influence the reaction time negatively. The correlation between preactivation level and tCNV amplitudes (r = 0.26, P = 0.050) and reaction time and tCNV amplitudes (r = 0.25, P = 0.057) was close to significant. The increased preactivation level of migraine patients is reflected in the tCNV amplitudes. The tCNV amplitude affects the reaction time. Further research has found this relationship between tCNV, which represents motor preparation, and reaction time (20, 35).

The study has confirmed an effect of pregnancy on slow cortical potentials and reaction time in migraine. In this context, the preactivation level seems to play an important role. Further research with blood samples, measurement of steroid hormones and psychological factors in pregnancy are needed.

Competing interests

None to declare.

Acknowledgements

The writing of this manuscript was improved by the assistance of one reviewer, N. Birbaumer. We also thank Dr Michael Siniatchkin for helpful discussion.

References

Lance

Anthony

Some clinical aspects of migraine. A prospective survey of 500 patients. Arch Neurol 1966; 15:356–61.

Callaghan

The migraine syndrome in pregnancy. Neurology 1968; 18:197–9.

Somerville

BW.

A study of migraine in pregnancy. Neurology 1972; 22:824–8.

Rasmussen

BK.

Migraine and tension-type headache in a general population: precipitating factors, female hormones, sleep pattern and relation to lifestyle. Pain 1993; 53:65–72.

Granella

Sances

Zanferrari

Costa

Martignoni

Manzoni

GC.

Migraine without aura und reproductive life events: a clinical epidemiological study in 1300 women. Headache 1993; 33:385–9.

Cupini

Matteis

Troisi

Calabresi

Bernardi

Silvestrini

Sex-hormone-related events in migrainous females. A clinical comparative study between migraine with aura and migraine without aura. Cephalalgia 1995; 15:140–4.

Maggioni

Alessi

Maggino

Zanchin

Headache during pregnancy. Cephalalgia 1997; 17:765–9.

Granella

Sances

Pucci

Nappi

Ghiotto

Napp

Migraine with aura and reproductive life events: a case control study. Cephalalgia 2000; 20:701–7.

Scharff

Marcus

Turk

DC.

Headache during pregnancy and in the post partum: a prospective study. Headache 1997; 37:203–10.

10.

Marcus

Scharff

Turk

Longitudinal prospective study of headache during pregnancy and post partum. Headache 1999; 39:625–32.

11.

Chen

Leviton

Headache recurrence in pregnant women with migraine. Headache 1994; 34:107–10.

12.

Sances

Granella

Nappi

Fignon

Ghiotto

Polatti

Nappi

Course of migraine during pregnancy and post partum: a prospective study. Cephalalgia 2003; 23:197–205.

13.

Welch

Ramadan

NM.

Mitochondria, magnesium and migraine. J Neurol Sci 1995; 134:9–14.

14.

Ambrosini

Schoenen

The electrophysiology of migraine. Curr Opin Neurol 2003; 16:327–31.

15.

Schoenen

Deficient habituation of evoked cortical potentials in migraine: a link between brain biology, behavior and trigeminovascular activation?

Biomed Pharmacother 1996; 50:71–80.

16.

Lodi

Montagna

Soriani

Iotti

Arnaldi

Cortelli

. Deficit of brain and skeletal muscle bioenergetics and low brain magnesium in juvenile migraine: an in vivo 31P magnetic resonance spectroscopy interictal study. Pediatr Res 1997; 42:866–71.

17.

Bohotin

Fumal

Vandenheede

Gérard

Bohotin

Maertens de Noordhout

Schoenen

Effects of repetitive transcranial magnetic stimulation on visual evoked potentials in migraine. Brain 2002; 125:912–22.

18.

Walter

Cooper

Aldridge

McCallum

Winter

AL.

Contingent negative variation: an electric sign of sensorimotor association and expectancy in the human brain. Nature 1964; 25:380–4.

19.

Maertens de Noordhout

Timsit-Berthier

Timsit

Schoenen

Contingent negative variation in headache. Ann Neurol 1986; 19:78–80.

20.

Birbaumer

Elbert

Canavan

Rockstroh

Slow potentials of the cerebral cortex and behavior. Physiol Rev 1990; 70:1–41.

21.

Schoenen

Timsit-Berthier

Contingent negative variation: methods and potential interest in headache. Cephalalgia 1993; 13:28–32.

22.

Kropp

Gerber

WD.

Contingent negative variation during migraine attack and interval: evidence for normalization of slow cortical potentials during the attack. Cephalalgia 1995; 15:123–8.

23.

Kropp

Mueller

Gerber

WD.

Long-lasting migraine alters amplitudes and habituation of contingent negative variation. Cephalalgia 2007; 27:666.

24.

Siniatchkin

Andrasik

Kropp

Niederberger

Strenge

Averkina

. Central mechanisms of controlled-release metoprolol in migraine: a double-blind, placebo-controlled study. Cephalalgia 2007; 27:1024–32.

25.

25 Headache Classification Committee of the International Headache Society. The International Classification of Headache Disorders, 2nd edn. Cephalagia 2004; 24 (Suppl.1):1–160.

26.

Tfelt-Hansen

Block

Dahlof

Diener

Ferrari

Goadsby

. Guidelines for controlled trials of drugs in migraine, 2nd edn. Cephalalgia 2000; 20:765–86.

27.

Kropp

Siniatchkin

Gerber

W-D.

Contingent negative variation as indicator of duration of migraine disease. Funct Neurol 2000; 15 (Suppl.3):78–81.

28.

Böcker

Timsit-Berthier

Schoenen

Brunia

CH.

Contingent negative variation in migraine. Headache 1990; 30:604–9.

29.

29 Baos

Ester

Castellanos

Nocea

Caloto

Gerth

I-Max Study Group . Use of a structured migraine diary improves patient and physician communication about migraine disability and treatment outcomes. Int J Clin Pract 2005; 59:281–6.

30.

Buckwalter

Stanczyk

McCleary

Bluestein

Buckwalter

Rankin

. Pregnancy, the postpartum, and steroid hormones: effects on cognition and mood. Psychoneuroendocrinology 1999; 24:69–84.

31.

de Groot

Vuurman

Hornstra

Jolles

Differences in cognitive performance during pregnancy and early motherhood. Psychol Med 2006; 36:1023–32.

32.

Wilder

JF.

Stimulus and response: the law of initial value. Bristol: Wright 1967.

33.

Knott

Irwin

DA.

Anxiety, stress, and the contingent negative variation. Arch Gen Psychiatry 1973; 29:538–41.

34.

Annovazzi

Colombo

Bernasconi

Schiatti

Comi

Leocani

Cortical function abnormalities in migraine: neurophysiological and neuropsychological evidence from reaction times and event-related potentials to the Stroop test. Neurol Sci 2004; 25 (Suppl.3):285–7.

35.

Kropp

Kiewitt

Göbel

Vetter

Gerber

WD.

Reliability and stability of contingent negative variation. Appl Psychophysiol Biofeedback 2000; 25:33–41.

Effects of Pregnancy on Slow Cortical Potentials in Migraine Patients and Healthy Controls

Abstract

Keywords

Introduction

Materials and methods

Patients and participants

EEG procedure

Statistical analysis

Results

Clinical data: course of migraine days before pregnancy, during the third trimester and 4 weeks after delivery

Electrophysiological data: effects of pregnancy and migraine on slow cortical potentials

Slow cortical potentials at the first measurement

Slow cortical potentials at the second measurement

Discussion

Clinical findings

Electrophysiological findings

Competing interests

Acknowledgements

References