Brainstem activation in cluster headache: An adaptive behavioural response?

Abstract

Introduction

The functional neuroimaging of headache patients has revolutionized our understanding of the pathophysiology of primary headaches, providing unique insights into these syndromes. Indeed, functional neuroimaging studies have shown the activation of specific brain structures, the brainstem in migraine and posterior hypothalamus in cluster headache (CH), as well as in other trigeminal autonomic cephalalgias. We describe the functional neuroimaging findings in a patient suffering from CH headache, investigated with functional magnetic resonance imaging (fMRI) during typical pain attacks.

Material and methods

Two typical, consecutive CH attacks were investigated by two fMRI imaging sessions on the same day. Both fMRI scans were performed at rest, during the CH attacks and the pain-free state induced by subcutaneous administration of sumatriptan.

Results

Significant activation of the bilateral red nucleus, ventral pons and trigeminal root entry zone ipsilaterally to the pain side was detected during the pain state, in addition to the hypothalamic region ipsilaterally to the pain side.

Conclusion

Being that such structures are mainly involved in motor function and reactive behaviour, their activation, in our hypothesis, may be linked to pain avoidance and may well represent a defence reaction in cluster headache, which is characterised by a "fight-or-flight" type behavioural pattern during pain attacks.

Introduction

The functional neuroimaging of headache patients has revolutionised our understanding of the pathophysiology of primary headaches, providing unique insights into these syndromes. Indeed, functional neuroimaging studies have shown the activation of specific brain structures, the brainstem in migraine and posterior hypothalamus in cluster headache (CH), as well as in other trigeminal autonomic cephalalgias (1). Such hypothalamic activation has been hypothesised to play a role in both triggering and terminating CH attacks. Moreover, it might also give rise to a central permissive state leading to attacks, or represent a response to pain driven by the first division of the trigeminal nerve (2). We describe the functional neuroimaging findings in a patient suffering from chronic CH investigated with functional magnetic resonance imaging (fMRI) during two typical pain attacks.

Subjects and methods

A 36-year-old male, suffering from left-side chronic CH, diagnosed two years previously according the 2004 classification criteria of the International Headache Society, was investigated with fMRI. He reported an average of two episodes per day at regular time intervals (14:00 and 23:00), lasting an average of 130 minutes without medication. The patient had neither cardiovascular/cerebrovascular risk factors, nor other comorbid pathologies. Clinical assessment and extensive morphological imaging revealed nothing relevant. Various prophylactic treatments, including verapamil (240 mg/day), topiramate (300 mg/day), lithium (900 mg/day) and prednisone (25–75 mg/day), had been administered without any benefit and therefore withdrawn at least two weeks before the study.

Study design

A prerequisite for patient enrolment in the study, established on the basis of medical history, was a dramatic response to subcutaneous sumatriptan administration. Written informed consent was obtained and the study was approved by the local ethics committee. Two typical, consecutive CH attacks were investigated by two fMRI imaging sessions on the same day. Both fMRI scans were performed at rest, during the CH attacks and the pain-free state induced by subcutaneous administration of 6 mg of sumatriptan after the onset of the attack.

A priori, the duration of the functional examination was established to be long enough to allow registration of the asymptomatic state, the onset of pain and the pain-free state. The patient was placed in the scanner 15 minutes (min) before the onset of the attack. When pain occurred he was to communicate the onset by pressing buttons on a non-magnetic button-box and the intensity by visual analogical scale (VAS = 10/10). After 5 min, in order to avoid movements of the patients due to the pain and to induce a pain-free state, sumatriptan was subcutaneously injected. The end of the attack, i.e. the pain-free state (VAS = 0, achieved within 10 minutes), was communicated using the same push-button controls.

Data acquisition

Data were acquired on a 1.5 T scanner (Maestro Symphony, Siemens). Functional datasets were obtained by means of a T2* weighted gradient echo planar imaging (EPI) pulse sequence (repetition time (TR)/ echo time (TE)/ flip angle (FA) 3000 ms/60 ms/30°, TR delay 20 ms, field of view (FOV) 20 cm, matrix 64 × 64, voxel 3.125 × 3.125 × 5 mm³, gap 1.25 mm). We acquired 20 oblique axial slices, parallel to the anterior-posterior commissure plane, extending from the foramen magnum to the vertex; 400 functional images were acquired. Anatomical datasets were obtained using a T1 magnetisation-prepared rapid gradient echo three-dimensional (MPRAGE 3D) pulse sequence (TR/TI/TE/FA 2160 ms/1,100 ms/3.93 ms/10°, voxel 1 × 1 × 1.4 mm³, gap 0.7 mm); 160 oblique sagittal slices were acquired.

Pre-processing of MRI data

Data analysis was carried out using BrainVoyager QX 1.8 software. First, functional data were processed in order to increase the signal-to-noise (SNR) ratio and reduce artefacts. We performed slice scan time correction and 3D motion correction by means of rigid body transformations (the estimated displacement in each direction was less than 3 voxels) and spatial smoothing with a Gaussian kernel of 4 mm full-width-half-maximum (FWHM). Anatomical data were re-sampled to obtain isometric voxels of 1 × 1 × 1  mm³ then normalised to the Talairach and Tournaux standard space (3) to allow intersession statistical analysis. Functional and anatomical datasets were co-registered by translation, rotation and zoom operations on the x, y and z axes. Condition-specific effects (pain vs. no pain) were estimated using a boxcar approach connected with the hemodynamic response function. Processed functional data obtained from the patient underwent single-subject and multi-subject general linear model (GLM) analysis. A statistical map was generated as t contrast with a threshold value of p < 0.001. Activation was considered significant at a threshold of 0.05 corrected for multiple comparisons.

Results

When the pain state was compared to the pain-free state in both attacks (single- and multi-subject GLM study), a significant cerebral activation (p < 0.05, corrected for multiple comparisons) of the hypothalamic region ipsilaterally to the pain side was observed, as documented in previous functional studies (4–7). Moreover, activation was observed also in correspondence with the ipsilateral trigeminal root entry zone, the bilateral red nucleus and ventral pons without lateralisation (Figure 1A–1D). The coordinates (hypothalamus, red nucleus, ventral pons) and the Z score of the activation zones are listed in the Table 1. Figure 2 shows the temporal course of the blood oxygen level-dependent (BOLD) signals in the ventral pons in multi-subject GLM. Additionally, with a less conservative threshold of p < 0.001 uncorrected, trends of activation were observed in the pre-frontal cortex, anterior cingulate cortex, contralateral thalamus, ipsilateral basal ganglia and bilaterally in the insula and the cerebellar hemispheres.

Figure 1.

(A–D) Functional MRI activation pattern during cluster headache attack. Multi-subject GLM analysis shows cerebral activation (p < 0·001 uncorrected for descriptive purposes) areas in the posterior grey hypothalamus (A, axial view) ipsilateral to the pain side, in left trigeminal root nerve (B, coronal view), in ventral pons (C, sagittal view) and in bilateral red nuclei (D, axial view). The cerebral activation was superimposed on a T1-weighted anatomical reference.

Figure 2.

BOLD signals in the ventral pons (Y axis). Black zone (A): asymptomatic state. Green zone (B): onset of the headache attack and state of pain. Blue zone (C): administration of sumatriptan. Grey zone (D): pain-free state. On the X axis, the number of GE-EPI acquisitions (400) is indicated.

Table 1.

BOLD-signal increases during spontaneous CH attack, compared with the pain-free state (p < 0.05, corrected for multiple comparisons — multi-subject study).

	Talairach coordinates			Z score
	x	y	z	p < 0.05 (corrected)
Hypothalamus	–3	–3	–8	4.91
Red nucleus	±5	–18	–4	4.83
Ventral pons	0	–16	–20	4.42

BOLD: blood oxygen level-dependent; CH: cluster headache.

Discussion

Not only do our findings reproduce previous observations (4–7) of the activation of the hypothalamic region ipsilaterally to the pain side in CH, but they also provide new evidence that other specific brainstem structures may be involved in CH pathophysiology. The question arises whether the activation of the red nucleus and ventral pons is pivotal or epiphenomenal in the pathophysiology of the pain attack.

The red nucleus represents an important subcortical relay station of a massive primitive descending motor tract (the rubrospinal tract). Along with its well-established role in the motor system, functional studies suggest that it is involved in pain processing and aversive events (8,9). Indeed, the red nucleus receives inputs from the periaqueductal grey and sends projections to the lateral reticular nucleus and the intermediate region of the spinal cord, implicated in the processing of noxious stimuli. Its activation has been reported in visually triggered migraine (10) and in association with laser-evoked pain (8).

To the best of our knowledge, only one study (11) has demonstrated the activation of the ipsilateral ventrolateral midbrain, which extended over the red nucleus, in association with pain in hemicrania continua, whose clinical phenotype overlaps with that of migraine and trigeminal autonomic headaches. Furthermore, in paroxysmal hemicrania, a positron-emission tomography (PET) study documented the activation of the contralateral ventral midbrain, which straddled the red nucleus and substantia nigra (12).

The area of the ventral pons activated in our study corresponds anatomically to the pontine nuclei (along the ponto-cerebellar pathway) and the pyramidal tract. Although pons activation in primary headaches has been described in migraine (13) and hemicrania continua (11), it was in the dorsal rostral area. There, ventral pons and the red nucleus are well known to be mainly involved in motor function and, likely, in reactive behaviour. We hypothesise that the activation of the bilateral red nucleus and ventral pons we observed may be linked to the motor restlessness during the pain attack and to a sort of “fight-or-flight” reaction in CH, characterised by a typical behavioural pattern.

Indeed, a review (14) of international literature reporting the behavioural, humoural and functional imaging aspects of primary headaches shows that migraine and CH have different behavioural patterns during attacks. It has been postulated that the behavioural responses to migraine and CH pain are evolutionary conserved reactions consistent with sickness behaviour and defence reaction, respectively (14). The sickness behaviour observed during migraine attacks is a pan-mammalian adaptive response to internal and external stressors, characterised by withdrawal and motor quiescence, sympathoinhibition and lethargy, in which visceral pain signals a homeostatic imbalance of the body and/or brain. In contrast, the defence reaction in CH is of a fight-or-flight nature, with motor restlessness and agitation, in which pain is of the exteroceptive type.

We therefore hypothesise that the activation of bilateral red nucleus and ventral pons observed in this study may well be related to pain avoidance and a defence reaction in the behavioural pattern of CH and, likely, of trigeminal autonomic headaches.

The other noteworthy finding in our study is the activation of the ipsilateral trigeminal root entry zone. Surprisingly, the trigeminal nucleus activation, although expected, with the relaying of nociceptive information in the sensory pathway, has not been reported by functional imaging study in primary headaches to date. Only the ponto-medullary junction activation has been detected in hemicrania continua and was hypothesised to reflect trigeminal nucleus activation (11).

Consequently, the trigeminal root entry zone activation detected during CH attacks in our patient might be epiphenomenal of the common final pathway of pain driven by the trigeminal nerve.

However, although these preliminary findings are intriguing, further studies are required to confirm the activation of the red nucleus and ventral pons and the trigeminal region in CH and to better clarify its complex, emerging pathophysiology.

Clinical implications

A 36-year-old male, suffering from chronic cluster headache (CH), was investigated with functional magnetic resonance imaging (fMRI) at rest and during the CH attacks and the pain-free state induced by subcutaneous sumatriptan administration. When the pain state was compared to the pain-free state, a significant cerebral activation of the hypothalamic region ipsilaterally to the pain side was observed, along with the areas (prefrontal cortex, basal ganglia, thalamus, cingulate cortex, insula and cerebellum) known to be involved in the “neuro-pain matrix”. Moreover, activation was also detected in correspondence with the bilateral red nucleus and ventral pons and the ipsilateral trigeminal root entry zone.

The area of the ventral pons activated in our study corresponds anatomically to the pontine nuclei and the pyramidal tract. Ventral pons and the red nucleus are well known to be mainly involved in motor function and, likely, in reactive behaviour. We hypothesise that the activation of the bilateral red nucleus and ventral pons may be linked to the motor restlessness during the pain attack and to a sort of “fight-or-flight” reaction in CH, characterised by a typical behavioural pattern. As a matter of fact, it has been postulated that the behavioural responses to migraine and CH pain are evolutionary conserved reactions consistent with sickness behaviour and defence reaction, respectively.

Although these preliminary findings may highlight some aspects of the complex, emerging pathophysiology of CH in a behavioural perspective, further studies are required to confirm the activation of the red nucleus and ventral pons in CH and to better clarify its role, pivotal or epiphenomenal.