Sounding out migraine-related interactions between the brainstem and cortex

The clinical features of a migraine attack include symptoms generated by both the brainstem and the cortex. Significant changes in the activity of both brain regions during migraine have been clearly demonstrated with functional imaging techniques (1,2). How these brain regions interact before, during, and after a migraine attack remains much less clear. Cortical spreading depression (CSD), the slowly propagated wave of cortical activity that is believed to be the physiological substrate of migraine aura, can activate brainstem nociceptive pathways in animal models (3–6). It is well known that different types of brain injury can trigger CSD (7), but the trigger(s) for CSD in the setting of migraine are unknown. The paper by Vinogradova in this issue of Cephalalgia uses a novel approach, the audiogenic seizure model in rats, to show that CSD can be triggered by sound-evoked brainstem activity. These findings raise important questions about the potential role of the brainstem as a driver of changes in cortical activity during migraine.

Audiogenic seizures are brainstem seizures evoked by loud sound (8). They occur primarily in rodents and there are certain strains of rodents, some with identified genetic alterations (9), that have a marked propensity to develop this type of seizure. Audiogenic seizures are characterized by “wild running” behavior followed in some cases by tonic or tonic-clonic motor behavior. They are believed to be triggered by activity in the inferior colliculus, and they have been reported to be associated with electrographic seizure activity in the midbrain in periaqueductal gray and substantia nigra, as well as in the medulla, and lateral geniculate bodies (8).

The study by Vinogradova shows that repetitive but not single audiogenic seizures in rat, identified based on characteristic running behavior, are followed by unilateral CSD as measured by implanted electrodes. With further stimulation at higher amplitude, cortical seizures may also occur following the spreading depression event, and the CSD events may be bilateral. These findings are interesting on multiple levels. First, they suggest that there may be mechanisms by which activity in the brainstem could trigger CSD. Second, they represent an example of migraine-related changes in cortical activity triggered by a sensory stimulus. Finally, they may provide some new insights into the overlapping mechanisms of migraine and seizures.

A number of studies using immunohistochemical and electrophysiological techniques in rodents have shown that CSD can activate second-order trigeminal neurons via the trigeminal nerve (3,4). Others have provided strong evidence that CSD can modulate the activity of second-order trigeminal neurons directly by descending central pathways that are distinct from the trigeminal nerve (5,6). It has been postulated that these peripheral and central descending pathways mediate the activation of nociceptive pathways by CSD, thereby linking CSD (and aura) to migraine headache. The results presented by Vinogradova raise the possibility that there could also be ascending pathways from the brainstem to the cortex that are involved in the initiation of CSD.

In addition to direct ascending neural pathways, another possibility is that CSD could be evoked by vascular responses to audiogenic seizures. Goadsby and colleagues found that brainstem stimulation, specifically stimulation of the locus coeruleus, can produce substantial cortical vasoconstriction in multiple animal models (10,11). Prolonged sound exposure in rats may also be associated with decreased cortical blood flow, in addition to substantial increases in brain capillary permeability as well as epidural and ventricular hemorrhages (8).

These observations suggest that reduced blood flow, increases in capillary permeability, and even hemorrhages could be mechanisms by which brainstem activation could trigger CSD. It is interesting to note that lower amplitude stimulation evokes a unilateral CSD event. Previous work by the author indicates that this may be due to a laterality of the brainstem seizure event (12). The mechanisms underlying this laterality remain unclear, but these results indicate that an event triggered by a bilateral sensory input can provoke a unilateral cortical response.

The fact that in this model CSD can be evoked (albeit indirectly) by a sensory stimulus could have relevance to the clinical features of migraine. Many migraine patients report that their attacks are triggered by bright lights, loud sounds, or strong smells. Whether this sensory stimulation truly plays a role in the initiation of a migraine attack is far from clear. For many patients, it is likely that what is perceived as a sensory trigger is instead a reflection of heightened sensory sensitivity that is part of the premonitory phase of a migraine attack. Nonetheless, it is not unreasonable to speculate that sensory stimuli could provoke changes in brainstem and cortical excitability that could in turn play a role in the initiation of a migraine attack. The model presented by Vinogradova provides an opportunity to better understand how sensory input could influence migraine-related interaction between the brainstem and cortex.

Migraine and epilepsy may have substantial clinical and mechanistic overlap (13). Headache, sometimes with migrainous features, is common following a seizure. Much less common is the reverse, where a seizure occurs as a part of a migraine attack. Each of the genes known to cause familial hemiplegic migraine may also be associated with different types of seizures. Increases in glutamate and K+ are believed to play key roles both in CSD and seizures (13,14). It is unclear, however, why in some cases these overlapping neurochemical changes and increases in cortical excitability lead to CSD, whereas in other cases they cause a seizure. The study by Vinogradova indicates that in the audiogenic seizure model, CSD occurs prior to cortical seizures with a lower threshold of activation. By contrast, when bilateral seizure activity occurs with higher amplitude stimulation, bilateral CSD events follow the seizure. Higher-resolution recordings of these phenomena could lead to new insight into the specific cellular events that initiate CSD as opposed to seizure activity.

One technical issue worth considering in this study is that the rats have implanted electrodes, which can by themselves act as sites of initiation for CSD. The author points out that CSD events consistently begin with a significant delay following the audiogenic seizure, suggesting that the CSD events do not begin at the electrode but rather travel from a different site to the electrode. Nonetheless, the possibility remains that electrode implantation is associated with small areas of injury that act as foci for induction of CSD. Non-invasive monitoring of CSD, such as can be achieved with optical imaging, represents an alternative approach that can circumvent questions regarding invasive electrode recording of CSD (14).

Despite the fact that there is no clear parallel to audiogenic seizures in humans, the phenomenon has nonetheless been useful as a model of basic seizures mechanisms, and as a platform for the investigation of potential therapies (8). Similarly, while audiogenic seizures may not be directly relevant to migraine, the individual features of the model may be tools for greater understanding of some of migraine’s anatomical and physiological features. Further studies of this model may provide new insight into relationships between brainstem and cortex in migraine, into distinct cellular mechanisms underlying CSD and seizures, and into the potential roles of sensory stimuli as migraine triggers.