What does EEG reveal about brainstem involvement in migraine with brainstem aura? Electrophysiological findings across clinical phases — A systematic review and pooled prevalence meta-analysis
TL;DR
Conventional scalp EEG is usually normal during both symptomatic and interictal phases of MBA, consistent with the available quantitative evidence, supporting its limited role in confirming the diagnosis or localizing the underlying neurophysiological process.
Abstract
Migraine with brainstem aura (MBA) is an uncommon migraine with aura subtype characterized by transient brainstem symptoms and, in some cases, decreased level of consciousness (DLOC). Although contemporary migraine models increasingly emphasize cortical spreading depression (CSD) as a central mechanism, the occurrence of impaired arousal raises the question of whether dysfunction of brainstem arousal networks also contributes to MBA pathophysiology. Electroencephalography (EEG) may provide important physiological insights into this question, yet EEG findings in MBA have never been systematically synthesized. We performed a systematic review to characterize phase-specific EEG abnormalities and to explore their implications for the relative contributions of brainstem and cortical dysfunction in MBA. A systematic search of PubMed, Embase, Scopus, the Cochrane Central Register of Controlled Trials (CENTRAL), ClinicalTrials.gov, the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP), and medRxiv was conducted from inception through March 31, 2026. Structured case series underwent quantitative synthesis, whereas case reports were analyzed qualitatively. EEG findings were categorized according to the migraine phase whenever sufficient clinical information was available and were further stratified according to the presence or absence of DLOC. Forty-two studies comprising 189 patients met the eligibility criteria, including three structured case series (112 patients) and 39 qualitative reports (77 patients). Among the three structured case series, only one reported EEG findings during symptomatic attacks, in which normal EEG findings were observed in 87.5% (7/8) of patients. Quantitative synthesis of the three structured case series demonstrated a low pooled prevalence of interictal EEG abnormalities (4.21%, 95% CI 0.75–9.50; I 2 = 0%), indicating that interictal EEG recordings were predominantly normal. In contrast, qualitative analysis demonstrated that attacks associated with DLOC most frequently exhibited generalized rhythmic delta activity (GRDA). Nevertheless, serial EEG recordings demonstrated continued evolution of slowing abnormalities beyond the symptomatic phase despite clinical recovery. Conventional scalp EEG is usually normal during both symptomatic and interictal phases of MBA, consistent with the available quantitative evidence, supporting its limited role in confirming the diagnosis or localizing the underlying neurophysiological process. Instead, its principal clinical value lies in excluding alternative neurological disorders. Based predominantly on qualitative case reports, GRDA was the most consistently reported EEG abnormality in MBA with DLOC and may be consistent with transient dysfunction involving bilateral brainstem–diencephalic arousal networks during these attacks. Future prospective studies incorporating advanced EEG techniques and multimodal neuroimaging are warranted to further clarify the pathophysiology of MBA.