From neuroimaging to the bedside: Take the sheets off the thalamus

The common migraine phenotypes, in their episodic forms with and without aura, are disorders that manifest with recurrent transient ictal dysfunctions without interictal sequelae. This helps us to understand, on the one hand, why the nature of migraine continues to be so elusive and why, on the other hand, it is one of the most studied pain syndromes. A complex mixture of genetics and environmental factors interact to determine migraine susceptibility, which may be conceptualized as a milieu of subtle factors that is able to cyclically ignite migraine by affecting the interaction between neuronal and vascular components within the brain. The logical consequence is that these elusive factors, if present, should also be detectable in the pain-free period as underlying dysfunctions. Many methods are available nowadays for assessing neural functions atraumatically in humans, contributing to the great advances made in understanding the pathophysiological facets of migraine. Since the most frequent aura symptoms are visual disturbances and because clinical hypersensitivity to environmental light stimuli, i.e. photophobia, is experienced by patients not just during the attack but even outside, several research groups are particularly focused on the study of the visual system in migraine. Above all, using electrophysiological techniques it was observed that migraineurs show an increase in cortical responsivity to any kind of sensory information, visually comprised, with respect to normal subjects (1). However, several relevant questions about cortical hyperresponsivity have not yet been answered, including: Could hyperresponsivity account for the peculiar migraineur propensity to develop attacks? What are the neuroanatomical correlates of cortical hyperresponsivity? Is there any relationship between interictal visual hyperresponsivity and the visual discomfort reported by patients? Could cortical responsivity differ between migraine with and without aura so that they should be considered as two separate entities?

The study by Datta et al. (2) published in the current issue of Cephalalgia tried to answer the last three questions by focusing on the neuroimaging correlate of visual cortical hyperresponsivity. To this purpose, they used multi-modality high-resolution (3 Tesla) functional magnetic resonance imaging (fMRI) to examine the hemodynamic brain response (BOLD) to light stimulation in a large cohort of migraine patients during the pain-free state, as well as in healthy controls. This study is distinct from previous reports on the same topic for three reasons: (I) the authors collected a considerably larger cohort of migraineurs and healthy subjects; (II) they studied patients with (MA) and without aura (MO) separately, not pooled together; (III) they also measured baseline metabolic activity, as reflected by resting cerebral blood perfusion. However, certain limitations of the present study should be acknowledged. Although patients were attack free before and during the scanning session, they were not checked for subsequent attacks that might have happened after the scan. This is a very important point, since from neurophysiological studies it emerges that migraineurs’ brain responses change with the time to the next attack as well as in the pre-ictal period, i.e. within 12–24 hours immediately preceding the attack (1). Another limitation, intrinsic in the methodology, is that even if blood flow is a useful surrogate marker for detecting pre-synaptic activity within brain areas, it does not allow the determination of whether the underlying physiological event is excitatory or inhibitory, or another energy-consuming process. Therefore, a direct comparison with results provided by electrophysiology is hard to make.

The authors found a significant BOLD amplitude enhancement in the occipital pole, corresponding to the foveal representation, in response to photic stimulation in MA patients, but not in MO patients, even if visual discomfort scores obtained in both patient groups were equally higher than in healthy subjects. However, another distinctive finding exclusively pertinent to MA patients is that the BOLD amplitude enhancement was not restricted to the primary visual cortex (V1), but was also present in the thalamic station of the visual pathway, the lateral geniculate nucleus (LGN). As baseline resting perfusion did not differ between groups, it could not explain the subsequent different activation in response to light.

Present data of increased hemodynamic activity in the LGN and V1 exclusively in MA patients supports the view that considers MA and MO as separate entities with peculiar hemodynamic and neurophysiological patterns. Moreover, this datum allows us to further underline the fact that the great heterogeneity of the clinical phenotypes of migraine is often underestimated. Despite a common diagnostic denominator, some clinical features such as the presence of prodromes, the type of aura symptoms (visual, somatosensory, dysphasic), the coexistence of MA and MO or associated symptoms such as vertigo, may characterize subgroups of patients subjected to different underlying pathophysiological and genetic mechanisms.

It is of interest that the data of Datta et al., confirming previous findings but in a larger patient population (3), did not find an association between light-evoked BOLD responses and visual discomfort. Therefore it should not definitively be excluded that visual pathways, and above all their thalamic relay, might be involved in the process leading to hypersensitivity to environmental light stimuli, that is, photophobia. Recent studies have begun to elucidate its anatomy and pathophysiology. In an animal model, the firing rate within the trigeminal nucleus caudalis neurons increased on light exposure, probably as a nociceptive response to light (4). Noseda and colleagues (5) identified a population of intrinsically photosensitive retinal ganglion cells (6) that make direct connections with lateral geniculate complex nuclei, which in turn receive inputs from trigeminal and retinal afferents, being thus unique in perceiving light as a nociceptive signal. Nonetheless, the authors found that these specific thalamic neurons project to multiple brain areas, including visual, somatosensory and associative cortices, suggesting a multisensory integrative response for photophobia. The thalamus seems to play a major role in this respect, not just as a relay station but also as a site of sensory integration (7). According to this view, the thalamic activation observed interictally in MA patients by Datta et al. might be considered an even more interesting finding than the generic V1 activation in response to light, the latter being also activated during a spontaneous migraine attack (8). Clear thalamic involvement in migraine pathophysiology was already disclosed in studies using refined neurophysiological electroencephalogram (EEG) and evoked potential (EP) techniques, which overall have indicated that migraine is characterized interictally by dysfunctional thalamo-cortical connections (9). The latter might account for altered cortical sensory information processing (10).

Future studies in the subject area in coming years may lead to better understanding of the mechanisms underlying the cortical hyperresponsivity in the different migraine phenotypes, its variations with the migraine cycle and its relation to changes in functional connectivity between the thalamus and cortex and in the activity of subcortico-(thalamo)cortical aminergic pathways. It will also be of the uttermost importance to gather more data on the hemodynamic and metabolic correlates of cortical hyperresponsivity, i.e. lack of habituation, by co-acquiring fMRI and evoked EEG. This would permit the bringing together of the excellent temporal resolution of EEG and the greater spatial precision offered by MRI.