Abstract
The aim was to study mismatch negativity features and habituation during the interictal phase of migraine. In migraine patients, a strong negative correlation has been found between the initial amplitude of long latency auditory-evoked potentials and their amplitude increase during subsequent averaging. We studied 12 outpatients with a diagnosis of migraine without aura recorded in a headache-free interval and 10 gender- and age-matched healthy volunteers not suffering from any recurrent headache. The experiment consisted of two sequential blocks of 2000 stimulations, during which 1800 (90%) recordings for standard tones and 200 (10%) for target tones were selected for averaging. The latency of the N1 component was significantly increased in migraine patients in respect of controls in both the first and second repetitions; the MMN latency was increased in the second repetition. In the control group the MMN amplitude decreased on average by 3.2 ± 1.4 μV in the second trial, whereas in migraine patients it showed a slight increase of 0.21 ± 0.11 μV in the second repetition. The MMN latency relieved in the second trial was significantly correlated with the duration of illness in the migraine patients (Spearman correlation coefficient: 0.69; P < 0.05). The increases in N1 latency and MMN latency and amplitude, the latter correlated with duration of illness, seemed to be due to a reduced anticipatory effect of stimulus repetition in migraine patients. This suggests that such hypo-activity of automatic cortical processes, subtending the discrimination of acoustic stimuli, may be a basic abnormality in migraine, developing in the course of the disease.
Introduction
Neurophysiological examination of the nonsymptomatic phase in patients suffering migraine detects an increased amplitude of evoked potentials or event-related potentials that is thought to reflect a condition of neuronal dysfunction that may predispose to the onset of the attack (1–3). Using visual, acoustic or event-related potentials in migraine, previous studies concur with the hypothesis that the high amplitude of averaged potentials may be caused by deficient habituation, or even potentiation, of the evoked response during stimulus repetition and increasing intensity (4–6). This might reflect a crucial abnormality in cortical information processing in that the deficient habituation might increase energy demands and play an important role in the pathogenesis of migraine (7, 8).
Symptoms of auditory dysfunction such as phonophobia are well known in migraine patients (9). In a recent evaluation of the headache-free phase of the disease, psychophysical discomfort thresholds to sound stimuli were also decreased in subjects compared with healthy controls and correlated with lower intensity sound levels capable of eliciting brain-stem auditory-evoked potential (BAEP) waves IV–V (10). In migraine patients, a strong negative correlation was found between the initial amplitude of long latency auditory-evoked potentials and their amplitude increases during subsequent averaging, suggesting that the response potentiation in migraine is likely caused by a reduced preactivation level of sensory auditory cortices (5).
Mismatch negativity (MMN), a particular type of event-related potential, is generated by the brain's automatic response to change in auditory stimulation (11). It can be measured in the absence of attention and without any task requirements, which makes it particularly suitable for testing different clinical populations; in this way, the analysis of MMN is believed to reflect the automatic central processing of a novel stimulus (12, 13).
MMN is a negative component usually peaking at 100–200 ms from the onset of any discernible change of the stimulus (‘deviant’) during a repetitive auditory stimulation (‘standard’). It is also assumed that this event-related potential may be the only objective measure for the accuracy of central auditory processing, the duration of echoic memory and of permanent auditory memory traces (14). The principal generators are located in the auditory and frontal cortices (15, 16). We studied the MMN features and habituation during the interictal phase of migraine to provide an objective measure of central auditory processing and to assess the automatic central elaboration of acoustic stimuli. To the best of our knowledge, the MMN paradigm has never before been applied to migraine sufferers.
Methods
Subjects
Twelve outpatients attending the Headache Centre of the Neurologic Clinic of Bari University, who fulfilled the criteria of migraine without aura, according to the International Headache Society (17), participated in the study. All patients underwent a standardized interview as well as clinical neurological and psychiatric examinations. They had been attending the practice for at least 12 months, during which they had been requested to register any headache episodes in a diary. Subjects with general neurological or psychiatric diseases, according to the Diagnostic and Statistical Manual of Mental Disorders (18), and any patients who were taking psychoactive drugs, prophylactic treatments for headache, or had displayed over-use of analgesic drugs in the last two months, were excluded. Only patients recorded as being in a headache-free interval (at least 72 h after the end of the last attack and 72 h before the next attack, the latter being verified by telephone interview) were considered for this study. Ten gender and age-matched healthy volunteers not suffering from any recurrent headache were recruited from the hospital/laboratory staff. All subjects were submitted to audiometric examination and only those with normal hearing functions were admitted to the examination. They all gave their informed consent to the study, which was approved by the Local Ethics Committee of our Department. The clinical features of the subjects are summarized in Table 1. Patients and controls did not differed significantly for age (results of
Clinical features of migraneurs
Stimuli and recordings
The subjects were seated comfortably in an armchair in a quiet, electrically shielded room. Tones of 40 ms duration for novel and 60 ms for frequent sounds, with a rise and fall time of 10 ms, were delivered binaurally through earphones. The intensity was 80 dB sound pressure level (SPL). The tone pitches were 1000 Hz for standard (i.e. frequent) stimuli (90%) and 2000 Hz for novel (i.e. rare) stimuli (10%). They were presented in a random order with a fixed interstimulus interval of 1 s. The experiment consisted of two sequential blocks of 2000 stimulations, during which 1800 nonartefacted recordings for standard and 200 for target tones were selected for averaging. The subjects were not informed about the occurrence of two different tones and were instructed to distract themselves by reading a book.
Electroencephalograph (EEG) recordings were obtained using a MICROMED (Mogliano Veneto, Italy) System 98 apparatus. Electrodes were positioned according to the international 10–20 system at positions Cz, Fz, F3 and F4 with reference to linked mastoids. Any EEG tracks contaminated by eye movements or other extra-cerebral artefacts producing voltage variation exceeding 10% of the maximal signal amplitude at any electrode were automatically omitted from averaging. A notch filter was also applied. The EEG signals were sampled at a rate of 250 Hz and filtered between 0.1 and 100 Hz.
Measurements
Evoked responses were analysed on the averaged responses, in terms of peak latencies and amplitudes of the maximal negative deflections within specified latency ranges determined by visual inspection. The N1 component was measured in the range of 50–150 ms post stimulus, for the MMN component the range of 100–200 ms from change onset was evaluated, according to Naatanen et al. (11). Habituation was assessed as the amplitude change in the second repetition block compared to the first repetition one. Due to the length of each trial, only two consecutive trials were performed for each subject. The examiner was blind regarding the subject's identity and diagnosis.
Because the obtained data were normally distributed (by Kolmogoroff-Smirnoff test) and characterized by homogenous variances (by F-test), the latencies and amplitudes of the evoked responses in each trial were compared by one-way
Results
The N1 component was clearly detected in response to both the frequent and the deviant stimulus in all subjects, whereas the MMN component was clearly evoked by the deviant stimulus. All the measures were taken on the CZ derivation, on the track resulting from the deviant tone. The latency of the N1 component was significantly increased in migraine patients in comparison with controls in both the first and second repetitions, and the MMN latency was increased in the second repetition (Fig. 1a). The results of one-way
Mean and standard deviations of (a) N1 and MMN latencies and (b) amplitude in migraine patients (□, n= 10) and controls (▪, n= 12). Results of one way
In the control group the MMN amplitude decreased on average by 3.2 ± 1.4 µV in the second trial, whereas in migraine patients it showed a slight increase of 0.21 ± 0.11 µV in the second repetition. The amplitude of MMN varied significantly in the two repetitions between patients and controls (
Discussion
A latency increase was detected in migraine patients for both the N1 and MMN components, the latter in the second repetition. This is the first time that the auditory N1 component has been shown to be delayed in migraine patients. In the most recent study on acoustic event-related potentials in migraine, no difference in N1 latency was detected between patients and controls (5). The length of our experiment, with many responses being recorded in a long time, may explain why this increase in N1 component latency has not previously been detected. In normal subjects, N1 latency actually became significantly shorter over time, reflecting increasing confidence in stimulus discrimination with repeated testing (19). The same phenomenon may have caused the lack of amplitude decrease with the stimulus repetition of other evoked responses. In this study, the N1 amplitude showed a trend towards a reduction in the second repetition that was similar in the two groups, as previously reported (4). However, the absolute N1 latency was further increased in migraine patients and slightly reduced in control subjects in the second trial, confirming a reduced anticipatory effect of stimulus repetition on N1 onset in our migraine series, which may also affect the N1 latency within each trial. The same pattern also seemed to affect the MMN latency, which appeared to be further increased in the second repetition in migraine patients. This differed from normal subjects, in which the second trial caused a latency reduction. Though a phenomenon of reduced anticipatory effect due to stimulus repetition might have emphasized a global increment of latency of N1 and MMN, especially in the second repetition, it was probably not strong enough to cause a significant difference in latency behaviour across the two repetitions between groups. Conversely, in our migraine subjects a significant lack of habituation affected the MMN amplitude. This peculiarity in the processing of repetitive information in migraine is similar to that reported for contingent negative variation (CNV) (20), visual evoked potentials (VEP) (21) and auditory evoked potentials (AEP) after high intensity (70 dB) stimulation (7). Reduced habituation in migraine is often described as a basic feature of the sufferer's brain (22) The lack of habituation of evoked potentials and ERPs may be subtended by the brain's inability to adjust the information processing to environmental demands; this leads to a neural overload and finally to a migraine attack (23).
In our series, the MMN amplitude started to reduce in the first repetition in normal subjects and tended to increase with the second repetition (Fig. 2). This seemed a compensatory phenomenon, confirming previous studies in which a strong negative correlation between the initial amplitude of evoked potentials and their amplitude increase during subsequent averaging suggested that the response potentiation in migraine is likely to be due to a reduced preactivation level of sensory cortices probably subtended by a reduced serotonergic central transmission (5). Subcortico-cortical monoaminergic pathways play a pivotal role in controlling cortical excitability; because of their diffuse innervation of sensory cortices and their ‘pacemaker’-like activity, serotonergic neurons in the raphe nuclei are particularly suited to modulate cortical information processing (24). In this way, the latency increase of N1 and MMN components emerging from long-duration trials might also support the hypothesis of a hypo-functioning state setting of cortical monoaminergic pathways (25). The latency increase of N1 and MMN and the MMN amplitude variation across the two repetitions seemed correlated phenomena, as shown by the significant results of Spearman tests. In addition, whereas the rate of MMN amplitude variation seemed independent of any clinical feature, the MMN latency increase was correlated with the duration of illness, suggesting that the hypo-functioning state of cortical automatic processing of acoustic stimuli may also develop in the course of the disease. The N1 and MMN waves are independent components, originating in different regions of the secondary acoustic cortex (26). In addition, there is also a frontal MMN generator (15). In our migraine patients, the slight abnormalities affecting the N1 latency may suggest a mild slowing in cortical processing of auditory stimuli, probably subtended by a slight reduction in adaptation to long-timing acoustic stimulation. Certainly, MMN appeared to be more involved by dishabituation phenomena in our migraine subjects; a number of studies have described changes in cognitive ERP in migraine and loss of acoustic P300 habituation has been reported in migraine (4). The MMN thus provides an objective evaluation of the accuracy of central auditory processing (13). From our results, we suggest that the dysfunction in cortical information processing in migraine, consisting in a basal hypo-cortical activity with a consequent reduced pattern of habituation and slowing of the cortical response linked with the deviant stimulus, may result in a subtle failure of the automatic processes of elaboration of acoustic inputs. This may be a feature intrinsic to migraine, as may be supposed based on the independence of reduced MMN habituation from illness severity. However, it may more likely result in a global slowing of auditory processes, developing in the course of the disease. In previous studies, subtle verbal dysfunction was detected in migraine sufferers, particularly on measures of verbal comprehension. Our results showing the slight abnormalities of the N1 and MMN pattern, consisting of both lack of habituation and latency increase, confirm that this cognitive impairment might be linked with the slowing and hypo-activity of automatic processes subtending the discrimination of acoustic stimuli (27). Further studies may be designed to specifically adapt the MMN method to explore auditory functions involved in verbal comprehension (27). The study of larger series of patients, including those with childhood migraine, and of the different phases of migraine, should help clarify how the possible constitutional abnormalities in cortical functioning and automatic cortical processing of sensitive and sensorial information in migraine patients may be modified in the course of the disease.
Illustrative recordings in two successive blocks of N1 and MMN components in a (a) control subject and (b) in a patient suffering from migraine without aura. Continuous lines represent ERPs elicited by deviant tones; Broken lines represent ERPs elicited by standard tones.