Whether migraine pathophysiology stems from vascular or centrally-driven origins has been debated for decades. However, facilitated by the development of modern neural imaging techniques and scientific technology, the last century has seen the largest advance in our understanding of migraine. It is now well accepted that sensitization of the trigeminovascular pathway plays a crucial role in the initiation and expression of a migraine. This is supported by experimental human studies that revealed abnormal activity of the trigeminovascular system. This abnormal activity was found particularly in areas of the brainstem, midbrain and hypothalamus during a migraine attack itself and during the interictal period, that is at least 72 hours following and not within 24 hours before a migraine. Research into the premonitory period, the critical 24-hour pain-free period preceding a migraine, is scarce and as a result, there is a gap in our understanding of how and why sensitization occurs. It may be that altered brain function, particularly in brainstem sites, may either trigger a migraine itself or facilitate a peripheral trigger that activates certain pain-processing brain regions, resulting in head pain. As it is impossible to predict when a migraine is imminent, few studies have investigated the premonitory period. Understanding the underlying mechanisms of the migraine cycle has potential to transform the way migraine disorder is treated. The aim of this thesis was to identify functional brain differences throughout the migraine cycle, in particular in the critical 24-hour pain-free period preceding a migraine.
The first investigation (Chapter 2) aimed to identify if neural activity within the brainstem and hypothalamus would alter over the migraine cycle. I employed high-resolution functional magnetic imaging (fMRI) to measure ongoing activity patterns reflected through infra-slow oscillations (ISOs) and functional connectivity in the interictal, postdrome and premonitory periods of migraine compared with controls. A comparison between all groups provided evidence of unique activity in the 24-hour period immediately preceding a migraine. Increased ISO activity occurred exclusively in this period in areas of the trigeminovascular system including the spinal trigeminal nucleus (SpV), midbrain periaqueductal gray (PAG), dorsal pons, thalamus and hypothalamus. Remarkably, midbrain and hypothalamic sites were found to display increased functional connectivity and regional homogeneity immediately preceding a migraine suggesting a role for the PAG-hypothalamic interaction in migraine expression. Importantly, interictal and postdrome groups displayed similar activity as control groups, highlighting the unique nature of the premonitory period. It is possible that these increases in ISO power and regional homogeneity result from enhanced amplitude and synchrony of oscillatory gliotransmitter release immediately before a migraine attack, thus supporting the role of astrocytes and gliotransmission in migraine initiation and/or expression. These findings have never been reported in the premonitory period of migraine and reflect altered brainstem and hypothalamic function immediately preceding a migraine.
Along with the central nervous system, cerebral vasculature changes have been strongly implicated as critical for migraine initiation. The second investigation (Chapter 3) aimed to build on my previous study by determining whether changes in absolute activity levels, reflected through abnormal cerebral blood flow (CBF), could be identified throughout the migraine cycle. I used pseudocontinuous arterial spin labelling (pcASL) to measure CBF in the interictal, postdrome and premonitory periods of migraine compared with controls. In line with the findings of my first investigation, this analysis revealed distinctive activity in the 24-hour period immediately preceding a migraine with decreased CBF in the hypothalamus, PAG and SpV. In addition, decreased CBF was revealed in higher brain structures such as the visual cortex, orbitofrontal cortex (OFC) and retrosplenial cortex. These findings also reflected alterations in the interictal group with decreases in CBF detected in higher brain structures including the nucleus accumbens, putamen, OFC and ventrolateral prefrontal cortex. Remarkably, decreased CBF in brainstem regions was found only in the period immediately preceding a migraine and these decreases occurred suddenly, as opposed to the decreased CBF found in the higher brain regions which tended to occur gradually throughout the interictal period as the migraine approached. The specialized activity of the brainstem in the period immediately preceding a migraine further emphasizes that brainstem abnormalities are involved in the initiation and/or expression of a migraine. Though many studies have explored CBF during other periods of migraine, this is the first study to measure resting CBF during the 24-hour period immediately preceding a migraine using ASL, and furthermore, to couple ongoing activity patterns (Chapter 2) with absolute brain activity.
The first two cross-sectional investigations (Chapters 2 and 3) revealed unique abnormal activity in the 24-hour period immediately preceding a migraine in areas of the brainstem, midbrain and hypothalamus. However, it remains unknown whether similar patterns would be revealed in a longitudinal study when comparing periods throughout an individual’s migraine cycle. To complete this thesis, the third investigation (Chapter 4) aimed to follow the migraine cycle of three migraineurs by imaging them five days a week over four weeks. Due to the cyclic nature of migraine, I expected that when comparing the activity in the 24-hour period immediately preceding a migraine with other periods of migraine within these individuals, the findings would reflect similar patterns as our cross-sectional studies. Indeed, using fMRI I explored resting brainstem activity patterns and found that although resting activity variability was similar in controls and migraineurs on most days, during the period immediately preceding a migraine, brainstem variability increased dramatically. These increases in resting variability were restricted to specific areas of the pain processing pathway including the SpV and dorsal pons. Remarkably, these changes were located in the same brainstem regions which have been shown to be activated during a migraine itself, but again they occurred whilst the individual was not in pain. These increases in brainstem variability were characterised by increased power at ISOs between 0.03-0.06 Hz and they were coupled to increases in resting regional homogeneity directly prior to a migraine. These oscillatory and regional homogeneity changes are consistent with the idea that changes in astrocyte function may precede a migraine and be responsible for its initiation and/or maintenance. These data provide the first evidence of altered brainstem function directly before a migraine throughout the migraine cycle of multiple individuals and provide compelling evidence for the hypothesis that brainstem function is altered immediately before a migraine.
Overall, these data reveal that the 24-hour period immediately preceding a migraine possesses unique qualities that may be crucial in the initiation and/or expression of the migraine. My findings reflect abnormal activity of the trigeminovascular system, in particular in areas of the brainstem, midbrain and hypothalamus. I found increases in ongoing activity patterns in the 24-hour period immediately preceding a migraine only, however abnormalities in absolute activity levels were also found in higher brain structures in the interictal period. Finally, when exploring the migraine cycle within three individuals, I found that the 24-hour period immediately preceding a migraine reflected very similar patterns to those revealed in my cross-sectional studies; relatively stable activity until the 24-hour period preceding a migraine, where a sudden over-exaggeration of activity occurred. It seems that migraine is indeed a cyclic disorder with brainstem function oscillating between altered states.