Cerebral blood flow alterations in migraine patients with and without aura: An arterial spin labeling study

Abstract

Abstract Background Migraine aura is a transient, fully reversible visual, sensory, or other central nervous system symptom that classically precedes migraine headache. This study aimed to investigate cerebral blood flow (CBF) alterations of migraine with aura patients (MwA) and without aura patients (MwoA) during inter-ictal periods, using arterial spin labeling (ASL). Methods We evaluated 88 migraine patients (32 MwA) and 44 healthy control subjects (HC) who underwent a three-dimensional pseudo-continuous ASL MRI scanning. Voxel-based comparison of normalized CBF was conducted between MwA and MwoA. The relationship between CBF variation and clinical scale assessment was further analyzed. The mean CBF values in brain regions showed significant differences were calculated and considered as imaging features. Based on these features, different machine learning–based models were established to differentiate MwA and MwoA under five-fold cross validation. The predictive ability of the optimal model was further tested in an independent sample of 30 migraine patients (10 MwA). Results In comparison to MwoA and HC, MwA exhibited higher CBF levels in the bilateral superior frontal gyrus, bilateral postcentral gyrus and cerebellum, and lower CBF levels in the bilateral middle frontal gyrus, thalamus and medioventral occipital cortex (all p values < 0.05). These variations were also significantly correlated with multiple clinical rating scales about headache severity, quality of life and emotion. On basis of these CBF features, the accuracies and areas under curve of the final model in the training and testing samples were 84.3% and 0.872, 83.3% and 0.860 in discriminating patients with and without aura, respectively. Conclusion In this study, CBF abnormalities of MwA were identified in multiple brain regions, which might help better understand migraine-stroke connection mechanisms and may guide patient-specific decision-making.

Keywords

• Migraine with aura
• Migraine without aura
• Arterial spin labeling
• Machine learning approach