Magnetic resonance spectroscopy during migraine attacks: A systematic review.

BackgroundThe neurobiological basis of migraine remains incompletely understood. Magnetic resonance spectroscopy (MRS) allows non-invasive quantification of neurochemical and metabolic patterns in the brain, offering unique insights into biochemical processes during distinct migraine phases. This systematic review provides a critical appraisal of existing evidence describing MRS-derived neurochemical and metabolic alterations during spontaneous and experimentally provoked migraine attacks.MethodsA systematic review was conducted in accordance with the PRISMA statement and prospectively registered in PROSPERO. Comprehensive searches of PubMed, Embase, and Scopus were performed from database inception through August 1, 2025. Eligible studies included observational or interventional investigations acquiring 1H-MRS or 31P-MRS data during the ictal phase in adults with migraine, incorporating either non-ictal comparisons or healthy controls. Considerable variability in study design, brain regions, and metabolite outcomes precluded quantitative synthesis, necessitating a structured qualitative analysis organized by MRS technique and anatomical region.ResultsEight studies published between 1988 and 2022 met inclusion criteria, comprising five 1H-MRS investigations and three 31P-MRS studies, some of which derived from overlapping participant cohorts. Brain regions examined included occipital cortex, pons, frontal cortex, basal ganglia, and parieto-occipital areas. In individual 1H-MRS studies, occipital cortex analyses demonstrated ictal elevations in total choline and total N-acetyl aspartate, while lower glutathione concentrations were observed. A single 1H-MRS study targeting the pons identified ictal increases in total creatine and total N-acetyl aspartate. Findings from 31P-MRS studies indicated altered cerebral energy metabolism during migraine attacks.ConclusionsAvailable MRS evidence suggests that migraine attacks are associated with altered cerebral energy metabolism, particularly within visual cortical and brainstem regions. However, existing studies differ substantially in design, acquisition parameters, regions of interest, and analytical approaches, such that few directly address comparable questions. Thus, the reproducibility of reported findings remains uncertain. Establishing reliable attack-related metabolic signatures will require well-designed longitudinal MRS investigations capable of directly probing ictal dynamics.