Abstract
The active 'oddball' event-related potential (ERP) P3 is elongated or reduced in migraine and tension-type headaches, indicating a deficit of active attention in these primary headaches. It is then reasonable to study the passive attention function in these headaches through the technology of passive paradigm singletone elicited ERPs. We invited 32 patients suffering from chronic tension-type headache (CTTH), 17 from frequent episodic tension-type headache (FETH) and 32 from interictal migraine without aura, as well as 28 healthy subjects to undergo passive paradigm single-tone ERPs. There were no statistically significant differences when the mean latencies and amplitudes of N1, P2, N2 or the mean latencies of P3 of the four groups were considered. In contrast, the P3 amplitudes were significantly reduced in the patient groups when compared with healthy controls. However, no further significant difference was found between patient groups when considering P3 amplitude. Our study demonstrated a deficit of passive attention in CTTH, FETH and migraine, but could not separate these headache types in this regard. The reduced P3 in patients might be due to the head pain experienced.
Introduction
During the cognitive process, attention is aroused where stimulus attracts our mental consideration, which includes stimulus orientation or selection, and mental resource utilization to detect the target and to block other irrelevant stimuli. In active or selective attention we wilfully direct our mental consideration at a stimulus, and in passive attention we involuntarily attend to the stimulus (1). Although studies concerning the cognitive functions in primary headaches such as migraine are far from conclusive (2), evidence has shown that attention problems exist both in migraine and in chronic tension-type headache (CTTH) (3, 4). The commonly used event-related potential (ERP), which indexes the development of attentional resources (5), provides an even clearer picture in this regard. For example, patients with migraine in the interictal period displayed the reduced P3 (P300 or P3b) component in the active ‘oddball’ paradigm in many studies and delayed P3 in some studies (6). Other investigations have shown normal active auditory P3 latencies in migraine and tension-type headache (TTH) (7), a prolonged visual P3 of the second trial in migraine (8) and a prolonged auditory P3 of the second trial in episodic tension-type headache (ETTH) (9). However, these results can not separate migraine from TTH.
Whether the P3 abnormality in these primary headaches was due to the head pain experienced remains unknown, since the chronic pain itself distorted the P3 component (10–12). Bearing in mind the deficit of active attention, one may ask whether or not passive attention functions normally in these headaches. There is an ERP technique which may help to address this issue, i.e. the passive paradigm single-tone elicited ERPs, during which process subjects are instructed to ignore the stimulus passively (13, 14). During the test procedure, the target tone occurs randomly in time as in the ‘oddball’ paradigm, but the standard tone is replaced by silence. Such a technique is simple and makes it easier to detect cognitive changes related to the involuntary orienting, i.e. passive attention. It has been documented that P3 obtained in this way shares similar latency and amplitude to those of the active ‘oddball’ P3 (13, 14).
Recent neuroimaging studies have shown that the external stimuli activate the human arousal state and elicit passive attention (15), therefore the high baseline arousal found in TTH as well as in migraine (16) might indicate a passive attention deficit in these primary headaches. On the other hand, based on previous findings that more stresses from common daily activities were perceived by patients suffering from CTTH compared with those from migraine (3, 17–21), we have hypothesized that the deficit of passive attention would be more severe in CTTH. In the current study design, the single-tone elicited P3 would be more reduced in this headache type than in migraine.
Methods
Subjects
This study was carried out on 109 subjects: 28 healthy subjects recruited from hospital staff or community (20 women; mean age 28.6 years with 10.7 SD, range 18–65 years), 32 patients with CTTH (12 women; 26.5 ± 7.6, 16–41 years), 17 frequent episodic tension-type headache (FETH; four women; 31.9 ± 10.1, 19–51 years) and 32 migraine without aura (MoA; 27 women; 32.4 ± 8.8, 17–53 years), who were diagnosed according to the classification criteria of the International Headache Society (ICHD-II) (22). Patients were verified to receive no prophylactic therapy and had to have been drug-free for at least 72 h. Migraine attack frequency was 1–6/month. The time interval between the last attack of migraine and the recording was at least 1 week. TTH sufferers were tested irrespective of head pain during the study procedure. Patients suffering from mixed headaches were excluded from the study. The time intervals between onset of head pain and the test day were between 3 and 360 months (mean 11.2 months). In addition, no subject had auditory system impairment or drug/alcohol abuse. There were no significant group differences when either age (one-way
Neurophysiological test
The subjects were seated comfortably in an armchair in a quiet room and were instructed to fix their eyes on a small piece of yellow sheet attached to the facing wall. Binaural tone stimuli (50 ms in duration with a rise–fall time of 10 ms) were delivered through earphones in a random series once every 2 s. The target tones (20%) were in 60 dB SPL and 2000 Hz, and the subject was instructed to ignore them passively; the standard tones (80%) were in 0 dB, i.e. whenever a standard tone in the ‘oddball’ paradigm would normally occur the subject heard nothing.
The ERPs were recorded with silver cup electrodes placed at midline Fz, Cz and Pz, and referenced to two linked cup electrodes placed on the earlobes, with a forearm ground. Bipolar recordings of electro-ocular activity (EOG) were made with electrodes placed at the outer canthus and supraorbitally to the right eye. The impedances of the electrodes were kept ≤5 kΩ. The potentials were amplified and averaged with a Nihon Kohden Neuropack-sigma device using a bandpass of 0.01–50 Hz. The configuration baud rate was set at 9600 bit/s and the data configuration length at 8 bits in the hardware of the device. The sampling pereiod was 100 ms prestimulus and 900 ms poststimulus. A sweep (about 10% in a total run) in which the EOG amplitude exceeded ±100 µV (indicating eye-blink or ocular movement) was excluded from averaging.
Thirty responses were analysed for each tone-elicited wave, in terms of peak latency and baseline-to-peak amplitude of the maximal negative or positive deflections determined by visual inspection. Latency ranges designated for potentials were: 70–160 ms for N1, 110–220 for P2, 150–300 for N2 and 200–450 for P3.
Statistical analysis
When treated as repeated measures, the potential component N1, P2, N2 and P3 mean latencies and amplitudes at three electrode sites in four groups were analysed by
Results
In compliance with the epidemiological findings (23), more migraine women and more TTH men were enrolled in the present study. Although there are some gender effects on the active paradigm ERPs (5, 24), the current study was focused on the group ERP differences. In general, our subjects showed clear N1–P2–N2–P3 component complexes at each electrode site. The grand mean of responses in patients with MoA, CTTH and FETH, as well as in healthy subjects are illustrated in Fig. 1. On average, all subjects displayed their P3 maximally at the centro-parietal zones.
Superi1mposed grand averages of event-related potentials in response to a passive auditory single stimulus (negative upward) and electro-ocular activity (EOG) traces recorded in healthy control subjects (—), chronic tension-type headache (……), frequent episodic tension-type headache (- - - - -) and migraine without aura () sufferers.
There were no statistically significant differences among subject groups in latencies of either the N1 (main effect, F[3,105] = 1.64, P = 0.18), P2 (F = 0.29, P = 0.83), N2 (F = 1.19, P = 0.32) or P3 (F = 1.08, P = 0.36) components. Nor were there any statistical differences among groups in amplitudes of N1 (F = 0.53, P = 0.66), P2 (F = 0.72, P = 0.54) or N2 (F = 1.61, P = 0.19) components (Table 1).
Latency and amplitude of event-related potentials in the healthy subjects, chronic (CTTH), frequent episodic tension-type headache (FETH) and migraine sufferers
P < 0.05 vs. healthy subjects.
By contrast, P3 amplitude was statistically significantly reduced in patients (main effect, F = 3.68, P = 0.015). Posthoc Duncan's test detected that at the three electrode sites, CTTH (Fz: P = 0.025; Cz: P = 0.014; Pz: P = 0.045), FETTH (Fz: P = 0.039; Cz: P = 0.025; Pz: P = 0.018) and migraine groups (Fz: P = 0.041; Cz: P = 0.031; Pz: P = 0.027) all displayed reduced P3 amplitudes when compared with the healthy control group. For example, the P3 amplitudes in CTTH were about 4 µV at Fz, 3.9 µV at Cz and 3.1 µV at Pz lower than those in the healthy controls; its amplitudes in FETH were about 3.5 µV at Fz, 3.3 µV at Cz and 3.9 µV at Pz lower; its amplitudes in migraine, on the other hand, were about 3.6 µV at Fz, 3.4 µV at Cz and 3.6 µV at Pz lower. However, there were no significant differences between patient groups when referring to the P3 amplitudes at any electrode.
When referring to the correlation between parameters, only P-values <0.001 were considered to be significant, to protect against Type I error. Consequently, there was no significant correlation found between subjects' ages, the educational level and ERP component latency or amplitude in any individual group. Nor was there any significant correlation found between headache duration and ERP component latency or amplitude in any patient group.
Discussion
All three patient groups displayed reduced P3 of the passive paradigm single-tone ERPs when compared with healthy controls, but there was no group difference between patient groups. Therefore, the deficit of the passive attention existed in patients suffering from CTTH, FETH and MoA. Although no previous study has demonstrated the passive attention deficit in primary headaches, our finding is in accordance with studies showing the reduced active ‘oddball’ P3 in migraine (6). In addition, following the first block of stimuli, the auditory novelty elicited P3 (or P3a), which indicates automatic attention switching, had a trend to be smaller (although non-significantly) in migraine sufferers when compared with healthy subjects (25). In order to avoid active attention involvement as much as possible, our test procedure was kept short in time, therefore it was not an ideal method to analyse the habituation of the passive P3 in the current study. However, another set of our very preliminary data indicated a trend of passive P3 dishabituation in migraine sufferers, but the lack of habituation was unrelated to their depressive moods.
Compared with its amplitude, the habituation dysfunction of P3 latency (i.e. successive prolongation) has been less reported in interictal migraine (6). In auditory modality, the latency change over successive trials was not very prominent. For example, the novelty tone elicited P3a latencies were about 229, 228, and 227 ms, respectively, in the first, second and third trials in migraine (25). The passive single-tone elicited P3 in migraine of the current study peaked at 282.7 ms at Fz, 280.0 ms at Cz and 283.6 ms at Pz. Nonetheless, the prolongation of the active oddball auditory P3 in the second trial (320.6 ms; the first, 313.6 ms) has been demonstrated in ETTH sufferers, but the dishabituation was not significantly different from healthy subjects (9), whereas the P3 latencies of FETH in the current study were 297.1 ms at Fz, 295.9 ms at Cz and 294.6 ms at Pz, respectively. Whether there is a P3 latency habituation disorder in migraine or in TTH remains to be answered.
Our study also favours a central dysfunction in these primary headaches. First, in migraine, the observed cognitive deficiencies have been extended to attention and memory, which accords with the proposal that migraine is a brain disease (26, 27). Second, the central problem has also been described in the pathophysiology of CTTH (28, 29). For example, altered pain perception and its detection have been reported in patients with CTTH (30, 31). The coping style through attention, and the retrospective recall of stressful daily events have also frequently been reported to be linked with the head pain itself or its exacerbation in patients with CTTH (3, 32). Third, the central mechanisms may contribute to the pathophysiology of FETH (33–35). Previous clinical evidence shows that the weekend as a trigger factor for migraine also plays a major role in patients with FETH (36). In an active ‘oddball’ ERP study, a remarkable increase of P3 latency has been demonstrated in these patients (9). It is possible that there is a conversion from FETH to CTTH, which might be an entirely central process (34).
In some cognitive studies in primary headaches, for example, disturbed visual attention was associated with the frequency of headache attacks and with head pain history (4). We failed to detect any correlation between the reduced P3 and headache duration either in migraine, in CTTH or in the FETH group. Whether the reduced P3 was due to the generalized condition, the head pain itself awaits further confirmation. Pain interrupts and demands attention, as some brain areas such as the anterior cingulate cortex mediate the impact of pain-related stress and allocate the attentional resources (37, 38). Both the experiment-induced pain in healthy subjects and the chronic pain in patients can decrease P3 amplitude (10–12).
From our limited neurophysiological results, we cannot discriminate migraine from CTTH or FETH, nor CTTH from FETH. This situation is similar to other studies which present a vague borderline between migraine and TTH. In some clinical studies, migraine, CTTH and FETH may share a common prodrome, aura sign or symptoms, as well as headache trigger factors (35, 39, 40). Some CTTH patients might have associating migraine symptoms and respond very well to sumatriptan treatment (41). Greater complexity comes with the psychological descriptions; the frequent reported symptoms of anxiety, depression or daily stresses in migraine also have been found in TTH (20, 42, 43). On the other hand, both the CTTH and FETH groups were recorded irrespective of head pain in our study and the results obtained were ambiguous with regard to discriminating migraine from TTH provided that these headaches were loaded on a continuum from ETTH/migraine to their chronic forms (33, 44, 45).
In conclusion, our study has shown decreased P3 amplitude of the passive paradigm single-tone elicited ERPs in patients suffering from MoA, CTTH and FETH, which might imply that there is deficit of passive attention in these primary headaches, and the P3 reduction might be due to the head pain experienced. On the other hand, our findings support the continuum hypothesis of migraine and TTH (33, 44). Nevertheless, one should note the limitations of our design, i.e. we recorded ERPs with only three midline scalp electrodes, which are less sensitive to detect the regional brain activity; we did not measure the depression or anxiety state in our subjects, while depression, for example, is often comorbid with migraine and CTTH (20, 42, 43) and influences the P3 morphology (46); moreover, we did not record the ERPs during the migraine attacks as a follow-up investigation, as there were dynamic changes of ERP components before and after attacks (6).
Acknowledgements
This study was supported by grants from the Department of Science and Technology of Zhejiang Province (No. 2005C33022), the Department of Health of Zhejiang Province (No. 2005QN011), the 151 Intelligence Fund of the Department of Personnel of Zhejiang Province, and the Research Centre of Language and Cognition of Zhejiang University to the corresponding author. W.C. and X.S. contributed equally to this study.