Studies of brain metabolism in migraine are limited, largely because of the inherent difficulties in applying complex methodologies to the study of a short-lasting paroxysmal disorder. Blood, platelets, muscle, and cerebrospinal fluid (CSF) have been examined for metabolic abnormalities, on the assumption that they may indirectly relate to metabolic changes in the central nervous system (CNS). Recently, functional imaging techniques have been employed: positron emission tomography (PET), magnetic resonance spectroscopy (MRS), and functional blood oxygen level-dependent magnetic resonance imaging (MRI-BOLD).

Skinhoj (34) documented significant elevation of lactate in the CSF during migraine attacks. CSF cyclic adenosine monophosphate (cAMP), a compound involved in cerebral energy metabolism, was found elevated during or within 48 hours of a migraine attack (38). Jain et al. (19) found magnesium (Mg²⁺) to be reduced in CSF sampled between migraine attacks. Shifts in CNS energy and electrolytic metabolism thus seem to take place during or in between migraine attacks. Limitations to the interpretation of these CSF findings include whether they were primary or secondary to the migraine attack. Moreover, either ischemia, which causes elevation of lactate and cAMP in CSF, or spreading depression, which raises glucose utilization, lactate, and cAMP, can account for the changes. Low tissue Mg²⁺ enhances the potential of CNS tissue to develop spreading depression (29). Elevated plasma levels of lactic and pyruvic acids in migraine patients suggests generalized impairment of energy metabolism (28), and assays of platelet respiratory chain enzymes in patients with migraine with (MA) and without aura (MO) detected decreased interictal activities of NADH-dehydrogenase, citrate synthase, and cytochrome-c-oxidase (31). These and similar findings in muscle (27) suggest a systemic impairment of oxidative mitochondrial function in migraine. Adenosine triphosphate (ATP) and its breakdown products AMP and adenosine perhaps released from “purinergic” nerves, were hypothesized to mediate vasodilatation and pain in migraine (9). Hyposecretion of ATP from platelet dense bodies (20) and 68% increased circulating adenosine levels during migraine attacks (14) suggest that migraineurs have abnormal purinergic metabolism.

Decreased cerebral blood flow (CBF) and oxygen extraction fraction (OEF) in the infarcted occipital cortex were observed in a patient with migrainous cerebral infarction studied by PET 14 days after occurrence (7) and in a patient with MA 90 minutes into an attack (17). Friberg et al. (13) used intra-arterial ¹³³Xe injection and arteriovenous oxygen difference sampling to measure CBF, oxygen consumption, and oxygen extraction in eight patients. Four developed attacks with aura and were found to show a 13% global increase in oxygen extraction, coinciding with a 12% drop in hemispheric CBF. Oxygen consumption and Paco₂ were unchanged. On the other hand, Andersson et al. (1) performed PET in migraine with and without aura induced by red wine. There was 23% reduction in CBF and 23% reduction in oxygen consumption in the primary visual cortex, but no change in oxygen extraction during the headache phase compared with baseline. Dissociation between abnormal CBF, metabolism, and neurologic dysfunction in migraine attacks was reported in a PET study in which bilateral occipital cortex oligemia was observed approximately 45 minutes before any complaint of visual symptoms (43). Finally, Bednarczyk et al. (5) found reduction in CBF and cerebral blood volume during an attack,
but no alterations in oxygen metabolism. Overall, the PET studies of oxygen metabolism, at this time, appear conflicting and do not permit confident interpretation of the findings.

Cerebral glucose consumption (CMRGL) was measured by 2-deoxyglucose PET in migraine patients before and after reserpine treatment (30). In patients given reserpine who experienced headache, CMRGL was globally reduced by 5 to 30% from baseline, in contrast to control subjects, who did not develop headache and had increased CMRGL after reserpine. These findings provide evidence for glucose hypometabolism occurring during migraine-like headache initiated by reserpine. The apparently contradictory findings from Mathew et al. (25), who found increased CMRGL in patients with aura status, could be reconciled by increased demand for substrate in brain tissue subjected to repeated depolarization and repolarization during repeated auras, and the hypometabolism found by Sachs et al. (30) could be explained by measurements being made in the late neuronally suppressed phase of spreading depression. Glucose hypometabolism in the affected frontobasal cortex, caudate, and thalamus was found in a patient with familial hemiplegic migraine (FHM) on the 6th day after an attack (15).

Functional MRI-BOLD measures relative changes in oxygenation of the brain circulation. On brain activation, an initial redundancy of flow with relatively less oxygen extraction causes hyperoxygenation of hemoglobin. This relative reduction in deoxyhemoglobin produces increased signal intensity on a T2-weighted MR image. With MRI-BOLD, neuronal activation and oxygenation state can be measured nearly continuously in a second-to-second time frame with millimeter resolution. Using this technique, strong evidence was obtained of increased BOLD signal reflecting hyperexcitability of visual cortex in association with visual dysfunction (18). In concert with this finding, in five MA and five MO patients, visual stimulation triggered suppression of initial activation after a mean of 4.3 minutes. This slowly propagated into contiguous occipital cortex at 3 to 6 mm/minute, and was followed by typical headache or visual changes at a mean of 7.3 minutes. Neuronal suppression was accompanied by hyperoxygenation of the occipital cortex during the early minutes of the attack (11). Hyperoxygenation of the occipital cortex was also substantiated in a patient with spontaneous visual aura: hyperoxia occurred in cortical gray matter only, not what would be expected during ischemia, and also in mesencephalic structures, despite the patient experiencing only left homonymous quadrantanopia (42). Indeed in MA patients, visual stimulation triggered typical headache and visual changes with onset coincident with hyperoxia and blood volume increase in the red nucleus and substantia nigra, preceding occipital cortex signal changes (10). Thus, headache and visual change in migraine are accompanied by spreading suppression of initial neuronal activation and increased occipital cortex oxygenation. Moreover, a neuronal network of brainstem structures appears activated during an attack (42). Because ischemia should decrease, not increase, T2-weighted image intensity, taken together with these findings, argue for a spreading depression like event.