Narrative Review of Neuroimaging in Migraine With Aura

Karissa N. Arca MD; Juliana H. VanderPluym MD; Rashmi B. Halker Singh MD


Headache. 2021;61(9):1324-1333. 

In This Article

Ictal Imaging

Noncontrast Computed Tomography of the Head and Magnetic Resonance Imaging of the Brain

Standard assessment for acute onset neurologic symptoms generally includes a noncontrast computed tomography (CT) of the head or magnetic resonance imaging (MRI) of the brain, usually as part of a stroke evaluation. A noncontrast CT of the head will be normal in a patient presenting with MA;[3] however, many patients experiencing acute ischemic stroke may also have normal findings on CT of the head. Some stroke centers have implemented MRI for evaluating the acute presentation of neurologic symptoms. Diffusion restriction on MRI is typically equated with early infarcted tissue and irreversible damage. Reversible, transient diffusion restriction without permanent tissue damage has been documented in seizure as well as migraine with aura.[4–6] In two case reports, diffusion restriction in MA was isolated to the splenium of the corpus callosum, an area vulnerable to "excitotoxic edema" due to elevated extracellular glutamate.[5,6] It is possible that the diffusion-weighted imaging abnormalities were related to CSD, which results in excess release of glutamate.[1]

Although focal diffusion restriction is not a common finding in MA, two case reports document cortical edema or cortical ischemic changes spanning multiple vascular territories in patients with prolonged aura or neurologic deficits with migraine.[7,8] In these cases, the abnormal diffusion restriction resolved on repeat imaging months after the event.[7,8] The prolonged symptoms and neuroimaging findings may have resulted from recurrent focal CSD and prolonged period of oligemia, which can cause migrainous infarction.[7] Recurrent spreading depolarization has been demonstrated in ischemic tissue, often occurring along the borders of the infarct and spreading into healthy tissue.[9] A prolonged oligemic phase can metabolically endanger tissue and may have been the nidus for recurrent focal CSD. One perfusion-weighted imaging study documented that blood flow to the affected area during MA was faster than blood flow during ischemic stroke, which highlights the importance of detailed vascular and perfusion imaging in addition to routine brain MRI for complex clinical scenarios.[10]

Reversible cortical venous engorgement on MRI in the symptomatic hemisphere has been documented in patients with acute neurologic changes later recognized as migraine with aura.[11–13] In oligemic conditions, such as the first phase of CSD, cortical neurons must extract additional oxygen to facilitate cellular metabolism, thereby resulting in elevated levels of deoxyhemoglobin recognized as prominent cortical veins on susceptibility-weighted imaging (SWI).[11] A case series of six patients who presented acutely with neurologic symptoms and had discharge diagnosis of MWA were found to have a single prominent cortical vein on SWI (the "index vein"), which correlated to the area responsible for the neurologic (aura) symptoms. Reversibility of this finding was demonstrated in one patient.[11] Reversible cortical venous engorgement has also been observed in the symptomatic hemisphere in case reports of typical and prolonged MA.[12,13] During an emergent evaluation for acute neurologic symptoms, hemispheric imaging changes that are not confined to a single vascular territory may be helpful in distinguishing stroke from MA. Recognizing the differences in neuroimaging findings in various neurologic emergencies may prove to be a valuable resource to determine the etiology of symptoms but does not appear to obviate the need for a thorough evaluation of other potential etiologies. Table 2 compares the acute neuroimaging findings in MA, ischemic stroke, and seizure.

Magnetic Resonance Angiogram (MRA) and Magnetic Resonance Perfusion

Ischemic stroke is typically determined by the presence of arterial stenosis or occlusion, and hypoperfusion in a single vascular territory. In contrast, vascular and perfusion changes in MA are not confined to a single vascular territory.[14] A case–control study investigated MRI findings within 3 h of acute neurologic symptom onset of 33 adult patients diagnosed with acute stroke and 33 with MWA.[10] Patients with stroke had findings restricted to a single vascular territory, whereas just over half of the patients experiencing MA demonstrated hypoperfusion in multiple vascular territories. Although most patients with MA had no vascular abnormalities on MRA, 15.2% had vasoconstriction in middle or posterior cerebral arteries, and 6.1% had mild dilation of these vessels.[10] A retrospective review of adolescents who presented to the emergency department and were discharged with diagnosis of MWA also found perfusion deficits contralateral to focal neurologic deficits spanning multiple vascular territories.[15] Seventy-five percent of the patients with hypoperfusion also had vasospasm in the same areas as the perfusion deficits.[15] The laterality of vasospasm and lack of beaded appearance or clinical presentation of thunderclap onset headache distinguished vasospasm from reversible cerebral vasoconstriction syndrome. Reported cases of MA attacks in adults do not seem to consistently demonstrate vasospasm that was found in pediatric cases.[10,16–18]

Like acute stroke, timing of imaging during MA can influence findings. Perfusion-weighted imaging during an MA attack has shown a pattern of hypoperfusion early in an attack followed by hyperperfusion, although on a limited case report and case series basis. A case series of four patients with MWA were imaged with arterial-spin labeling (a noncontrast option of perfusion imaging), two during aura and two in the early headache phase of migraine after aura resolution.[16] During aura, adjacent areas of hyperperfusion and hypoperfusion in the symptomatic brain area were demonstrated, whereas during the headache phase, only areas of hyperperfusion were present. The perfusion changes were not confined to a single vascular territory, and all patients also had normal imaging while asymptomatic.[16] The presence of hyperperfusion in the active headache phase may represent transient recovery of cerebral arteries to normal or a slightly dilated size as seen in animal models of CSD.[19] The exact physiology underlying the transition from hypoperfusion to hyperperfusion and whether it can be used as evidence for CSD is not clearly understood and has not been confirmed in larger sample sizes.[20]

The time it takes for cerebral perfusion to change during the aura or headache phase is another facet of MWA neuroimaging that requires further research. One case series reported hypoperfusion up to 6 h after onset of neurologic symptoms in one subject,[15] whereas others have documented the transition to hyperperfusion anywhere between 4 and 28 h on a case series and case report basis.[17,21] Once better understood, this "biphasic" evolution may be used to distinguish MWA from stroke or transient ischemic attack.[17]

Functional Imaging and Spectroscopy

Functional imaging performed during an MWA attack has been limited and its utility is research based rather than clinical. A historical study by Hadjikhani and colleagues observed three patients with near-continuous functional MRI (fMRI) during a visual aura attack. Early in aura, patients demonstrated focal increase in blood oxygenation level-dependent (BOLD) signal in the occipital lobe, which progressed rostrally over the occipital cortex, followed by a reduction in the BOLD signal. The progression across the occipital cortex corresponded with retinal topography of the visual aura. The BOLD signal is reflective of the equilibrium between oxygen transported to tissues and its utilization by the brain.[22] Therefore, the increased signal initially is thought to represent increased blood flow and neuronal activity followed by reduced signal due to oligemia and reduced neuronal metabolism.[22]