People that are interested in metabolic changes in migraineBased on the available literature ^1-3, altered cerebral blood flow (CBF) has been observed during migraine attacks and changes in cortical excitability have been described. Yet, it is unknown if these processes may be differentially involved in chronic and episodic forms of the disease.We examined 16 adult patients with episodic (EM, 2 males) and 21 with chronic migraine (CM, 6 males) as well as 26 healthy controls (HC, 10 males). We recorded attack frequency, pain intensity, aura occurrence and medication. To assess baseline CBF (during the interictal period), we used non-invasive arterial spin labeling magnetic resonance imaging (ASL-MRI). Specifically, CBF was recorded with a 2D pseudo-continuous ASL (pCASL) sequence ⁴ with background suppression (repetition time (TR)/echo time (TE) = 4200/16 ms, flip angle: 90°, voxel size: 3 x 3 mm, 20 slices, thickness: 6 mm, labeling duration: 1.65 s, post-labeling delay: 1.53 s) implemented on a Philips 3 Tesla Ingenia scanner. Equilibrium brain tissue magnetization (M0) images were recorded too. The ASL analysis was performed with the ASL toolbox ⁵. Pre-processing included motion correction, spatial smoothing and normalization. CBF quantification was done using the one-compartment model ⁶. CBF difference images were achieved by simple subtraction to minimize spurious BOLD contaminations within the CBF signal ⁷. We assessed main effects of group with an analysis of variance (ANOVA). Next, we calculated planned contrasts (t-tests, p < 0.001) between groups. For magnetic resonance spectroscopy (MRS) imaging, we used MEGA-PRESS ^{8, 9} to asses both GABA (inhibition) and Glx (combined glutamate and glutamine: excitation). MEGA editing was combined with the Point-Resolved Spectroscopy Sequence, inner volume saturation (Henning et al., 2008), and interleaved “VAPOR” water suppression ^{10, 11}. MEGA-PRESS with inner volume saturation increases sensitivity for edited single-voxel measurements of glutamate and GABA ⁸. We recorded ten blocks (320 averages total) and per block, the editing pulse was applied in an alternating manner either at 1.9 ppm or at 7.5 ppm. All blocks were retrospectively frequency aligned using the Choline and Creatine as frequency references. The radio frequency carrier frequency was set to Creatine. MRS parameters were: TR/TE: 1800 ms/68 ms, editing pulse duration: 16 ms. We recorded MRS from a voxel covering the bilateral visual cortex (30 x 30 x 30 mm), positioned on a T1-weighted 3D-MPRAGE image. The analysis was performed with TARQUIN (Wilson et al., 2011) that yields comparable reliability in metabolite concentration estimates to LCModel ^12-14. Spectral quality was assessed using Cramer-Rao lower bounds (CRLB); data with a CRLB > 20% were considered unreliable. The reference signal was set to H2O (= H2O scaled). Metabolite-to-H2O ratios were corrected for differences in grey- and white matter, and cerebrospinal fluid voxel content, influencing the measured H2O concentration ¹⁵. The resulting metabolite-to-H2O ratios are presented in institutional units (IU).Participants did not differ with respect to sex, age, and handedness. Pain scores (during MRI), depression and anxiety values were not different between the two patient groups. For CBF, a main effect of group was seen (F(3, 60) = 8.4, p < 0.001). EM revealed lower CBF than HC in fronto-central brain regions, including the striatum (Fig. 1A). In contrast, EM showed stronger CBF in the brainstem (Fig. 1B). The contrast ‘HC - CM’ yielded widespread hypoperfusion in frontal, somatosensory, parietal, temporal, cerebellar, and visual regions in CM (Fig 1C). EM demonstrated higher CBF than CM in visual, parietal, thalamic, cerebellar, temporal, orbito- and inferior frontal regions (Fig. 1D). Glx was reduced in CM (p = 0.002) and EM (p = 0.036) relative to HC (Figure 2). GABA was higher in the EM than CM group (p = 0.041). Headache days were not linked to levels of GABA or Glx within or across patient groups. CBF changes in migraineurs are observable in areas linked to pain processing such as the anterior cingulate and insular cortex, thalamus, and brainstem. CM showed even lower CBF relative to EM in visual areas, which might be linked to more severe signs of cortical spreading depression in chronically affected patients. Based on the MRS findings, we conclude that cortical excitability is altered during interictal periods in EM and CM, and that the excitation-inhibition balance is especially disturbed in CM, as not only Glx but also GABA is altered. However, it needs to be examined if it is glutamine and/or glutamate that are lowered in migraineurs.