Background
Migraine is the second most prevalent type of primary headache disorder, which affects approximately 14.7% of both males and females [
1]. Migraine is characterized by moderate to severe throbbing pain with sensitivity or intolerance to light and sound during the headache and is often accompanied by nausea and vomiting. Migraine, especially chronic migraine, is often comorbid with psychiatric disorders that exhibit affective temperament dysregulation and suicidal behaviors [
2,
3]. Migraine has been associated with an increased risk of vascular events, specifically cardiovascular disease [
4]. Vascular pathology is a strong risk factor for cognitive dysfunction [
5,
6].
In addition, migraine has been linked to an increased prevalence of clinically silent brain lesions and subtle gray matter damage which affect the cognitive processes [
7,
8]. These associations suggest that individuals with migraines have impaired cognitive function due to these structural lesions. However, previous studies of migraine and cognition have been contradictory. Some studies suggest that migraineurs have subtle interictal cognitive abnormalities aligning with attentional deficits [
9‐
11]. Some studies have reported improved performance [
12]. Some longitudinal studies have suggested fewer dysfunctions in the Mini Mental State Examination (MMSE) and executive functions in migraineurs. In addition, some studies have not found any significant differences in cognitive performance between migraine patients and controls [
13]. These inconsistencies could be due to methodological issues including different migraine assessment methods and the program of cognitive assessment.
Given the high prevalence of migraine in the general population, it is clear that the relationship between cognitive dysfunction and migraines present a significant public health interest.
Event-related potentials (ERPs) are one of the most useful tools in investigating neural substrates and cerebral regions involved in specific cognitive function because of the objectivity and noninvasiveness of the method. Undoubtedly, the P3 is the most studied cerebral wave used for evaluating cerebral information processing during the course of various neurological diseases because of its easy recording and reliability.
This study aimed to characterize cognitive testing and P3 in migraineurs. We hypothesized that migraine patients have cognitive dysfunction and P3 abnormalities, such as reduced P3 amplitude and/or a prolonged latency, suggesting alterations in the cognitive-evaluative component.
Discussion
The present study revealed that migraine patients had significantly lower scores than controls on cognitive tests and increased P3 latencies. Our findings suggest that migraineurs are associated with cognitive dysfunctions. Previously, Freitas and colleagues found that 49% of MoCA score variability was attributed to age and education [
15]. To exclude the influence of age and education on cognitive function, we had to balance age and education level between migraineurs and healthy subjects when we collected subjects. All participant ages were below 55 years old, and all had at least five years of primary education. For the psychological profile, migraineurs reported higher scores on the symptom checklist for anxiety. Some studies have shown that even subclinical levels of anxiety may negatively impact cognitive ability [
16,
17]. To eliminate the effect of psychology, Spearman’s correlations were calculated between cognitive tests and psychology tests, and we did not find any correlation between cognitive tests and severity of psychological impairment. The poorer cognitive performance exhibited by migraineurs was not determined by the emotional variables.
The present study revealed that migraineurs had significantly lower scores than healthy subjects on the delayed recall of ROCF test, the total MoCA and in five out of seven cognitive subdomains (i.e., language, executive functions, calculation, memory and orientation domains). The MoCA was included as a screening test for general cognitive abilities and out of a maximum score of 30 points, the cut-off value was 26 points [
18]. Eight migraineurs achieved scores below the available cut-off values in the MoCA test, but all of the scores were within two standard deviations of the mean. Four of them had low education, and two were elder than 50 years old. Most of the participants’ scores were not lower than the cut-off value reported in normative studies. Thus, we deemed that the reduced efficiency in selected cognitive domains did not correspond to a clinically relevant cognitive deterioration.
The longer durations and the increased frequency of migraine attacks were correlated with decreased cognitive performance. This indicates that the reduced scores on cognitive tests observed in migraineurs compared with controls is directly related to the length of time and the frequency of patient’s suffering from migraine.
Gil-Gouveia R aimed to document changes in cognitive performance of migraineurs under two conditions-during a naturally occurring untreated migraine attack and during a headache-free period. It was found that cognitive performance decreased during migraine attacks, especially in reading and processing speed, verbal memory and learning, supporting patients’ subjective complaints. Differences were found unrelated to pain intensity or duration of the attack [
19]. Riva D assessed the effect of migraine on the interictal functioning of children and adolescents and observed that both patient groups had cognitive performance within a normal range, except for a significant delay in the reaction time task. These researchers considered that slower reaction time to simple visual stimuli may be an early sign of a subclinical neuropsychological dysfunction, significantly correlated with the frequency of headache attacks [
20]. These data, together with our findings, would favor the view of some neuropsychological abnormalities in migraineurs as a consequence of the disease, or a perpetuating factor.
P3 is a generic name for a variety of relatively late, positive components with a centro-parietal midline distribution. Two principal electrophysiological markers have been considered as an objective index of cognitive processing: latency and amplitude. It is widely accepted that P3 latency reflects the length of stimulus evaluation processes [
21], when two choice reaction time (RT) is required and its amplitude is largely determined by stimulus relevance, the amount of attention allocated to the stimulus, working memory and the task’s complexity [
22,
23].
In our study, the latency for ERP components was significantly increased in migraine subjects. According to the previously mentioned facts, it could be suggested that migraineurs have a prolonged latency of P3, which represents a prolonged cognitive processing time. Some studies have shown that the amplitudes of P3 were significantly decreased in migraine patients compared with the healthy subjects, but the latencies of P3 did not show any significant effects [
24,
25]. Other results show significant elongation of latencies and a dysfunction in P3 amplitudes in the migrant group during headache attacks [
26]. However, several results have shown that migraine sufferers had longer P3 latencies, which is consistent with our findings.
The divergence among other studies and ours might be due to the different patient selection procedures. In our study, migraine patients were identified by expert neurologists according to established diagnostic criteria of the International Classification of Headache Disorders (ICHD-3 beta), while in previous studies, migraine patients were selected from a cohort of subjects on the basis of self-report measures. This procedure might have led to a misclassification of the non-migraine subjects. Furthermore, some articles have studied the cognition of children and adolescents. Cognition in children is immature, and we cannot compare it with adult cognitive function. Another potentially important factor is the procedure used for ERP. Differences in the tasks may probe different cognitive domains, and may lead to different conclusions.
In our study, ERP latency did not correlate with headache characteristics. This finding is not consistent with our expectations. ERP latency is a sensitive index, which is affected by many factors. The prolongation of latency may be affected by the migraine itself, but not by any characteristic of the headache.
In summary, the electrophysiological and behavioral data suggest that migraine is associated with cognitive dysfunctions, particularly, that cognitive dysfunctions can be related to the duration and the frequency of headache. Previously, people paid more attention to the pain of migraines and usually ignored the impairments of cognitive function in migraineurs. Although this functional phenomenon is present in the absence of clinically relevant deficits, it may reflect a vulnerability to executive high-demanding conditions of daily living activities in patients with migraine.
The strengths of this study include its standardized assessment of migraine status, information on migraine characteristics, and the availability of validated cognitive function measures. We discussed the correlations between neuropsychology, electrophysiology and migraine characteristics. We ruled out the effect of emotion on cognition. Our study had several limitations when interpreting our results. The study sample limited the ability to examine the effects of gender, age and education. Additionally, a previous study has shown that migraineurs with aura tend to exhibit even less decline over time than migraineurs overall [
27]. We did not find differences in mean cognitive scores between migraineurs with and without aura. However, this was likely to due to the sample sizes and resulting low statistical power. De Tommaso and colleges found that the changes in brain responsivity are associated with various stages of the migraine cycle, since migraine patients seem to have a response augmentation and poor habituation that normalizes just before/during attacks [
28]. In our study, patients were observed outside migraine attacks, but they could be at different points of a migraine cycle. We did not record the last time that they had a migraine attack. Furthermore, we did not observe the cognition of migraine patients at multiple time points, which limited us from examining the change in cognitive function over time.