• Continuous EEG detects nonconvulsive seizures in up to 18% of acute stroke. • Early EEG abnormalities predict post-stroke epilepsy risk and poor outcomes. • EEG biomarkers enhance post-stroke epilepsy prediction. Electroencephalography (EEG) can provide important diagnostic and prognostic information for the management of adult stroke patients. However, its integration into routine clinical practice remains variable. During the acute phase, routine and continuous EEG (cEEG) allow the detection of nonconvulsive seizures and status epilepticus (SE), especially in patients exhibiting impaired consciousness. Rhythmic and periodic EEG patterns, such as lateralized periodic discharges (LPDs), and early epileptiform abnormalities are associated with adverse outcomes and an increased risk of acute seizures. Quantitative EEG (QEEG) metrics, including alpha-delta ratio and asymmetry indices, have the potential to identify emerging ischemia and contribute to the prognostication of functional recovery. In the subacute and chronic stages, EEG assists in identifying patients at increased risk of post-stroke epilepsy (PSE). As such, EEG can enhance stratification and informs long-term management strategies alongside validated risk scores such as SeLECT, IsCHEMiA and CAVE. In the long term, serial EEG and QEEG assessments can guide the titration of antiseizure medications (ASMs), detect subclinical epileptiform activity associated with PSE and monitor recovery trajectories. Despite these contributions, evidence for EEG-guided interventions is limited, with variability in methodology, populations and protocols. Key gaps include lack of standardized strategies, limited outcome-related data and underrepresented groups. Future work should focus on multicenter trials, biomarker validation, and cost-effectiveness to clarify the role of EEG in stroke care. This review consolidates available evidence and highlights unmet needs in EEG use during stroke management, offering practical advice for clinicians and suggesting future research priorities.
Leclercq et al. (Wed,) studied this question.