Understanding epilepsy
An EEG records electrical activity from the scalp during the recording. It can support an epilepsy diagnosis, help classify seizure type or syndrome and sometimes capture an event — but it is not a pass-or-fail test for whether someone has epilepsy.
Central point: A normal EEG does not exclude epilepsy, and an abnormal EEG does not prove that every unexplained event is an epileptic seizure.
What an EEG actually records
Electroencephalography records voltage changes generated by large populations of cortical neurons and detected through electrodes placed on the scalp. A routine EEG usually samples a relatively short period of brain activity, so it may record normal background activity, nonspecific abnormalities, interictal epileptiform discharges, or occasionally an actual seizure.
Interictal epileptiform discharges
Spikes, sharp waves and spike-and-wave complexes can support a diagnosis of epilepsy when they fit the clinical context. They can also help distinguish focal from generalised epilepsies and contribute to syndrome classification. Interpretation belongs with the history, examination and other investigations; EEG findings should not be read in isolation.
Why a normal routine EEG does not rule epilepsy out
Epileptiform activity is intermittent. A person may have epilepsy but show no interictal epileptiform discharges during a short routine recording. NICE therefore explicitly states that EEG must not be used to exclude epilepsy.
Published yield varies by population and method. A population-based study found epileptiform abnormalities in 53% of people with incident epilepsy after the first EEG and 72% after the third. In a 2025 tertiary-centre study of newly diagnosed adults, the first routine EEG detected interictal epileptiform discharges in 31.6%, rising to 49.6% cumulatively after three EEGs. These figures illustrate why one normal test cannot settle the diagnosis.
Why a positive EEG is useful but not absolute proof
Epileptiform discharges are generally more specific than sensitive. A meta-analysis of routine EEG after a first unprovoked seizure found low sensitivity but high specificity in adults when epileptiform discharges were used as the positive finding. That means a clear epileptiform abnormality can substantially strengthen the diagnosis, while a negative recording is much less reassuring.
Incidental epileptiform abnormalities can occasionally occur in people without clinical epilepsy, and nonspecific slowing has many possible causes. The EEG report therefore has to be matched to the actual episodes being investigated.
Timing after a first seizure
NICE recommends that if an EEG is requested after a first seizure, it should be performed as soon as possible, ideally within 72 hours. Some observational studies have found a higher yield when EEG is obtained very soon after a seizure, although newer retrospective data have not always shown a statistically significant timing effect. The guideline recommendation therefore reflects the total evidence and clinical value of timely assessment rather than a guarantee that an early EEG will be abnormal.
Routine awake EEG
This is usually the first EEG used when epilepsy is suspected. The recording includes resting background activity and may include activation procedures such as hyperventilation and intermittent photic stimulation when appropriate and agreed. These procedures can increase the chance of revealing abnormalities in some epilepsy types, but their risks and benefits should be explained first.
Sleep and sleep-deprived EEG
Sleep can activate epileptiform discharges that are not seen while awake. If a routine EEG is normal but suspicion remains, NICE recommends considering sleep-deprived EEG after discussing the benefits and risks. Sleep deprivation itself can increase seizure risk in susceptible people, so instructions should be followed carefully.
Ambulatory EEG
Ambulatory EEG records for longer while the person continues much of their usual routine. NICE suggests considering ambulatory EEG, for up to 48 hours in its diagnostic pathway, when routine and sleep-deprived EEG are normal and diagnostic uncertainty persists. Longer recording increases the opportunity to sample sleep, interictal abnormalities and typical events.
Video-EEG monitoring
Video-EEG synchronises the electrical recording with video of the person’s behaviour. Capturing a typical event can be especially valuable when clinicians need to determine whether episodes are epileptic, functional/dissociative, syncopal, sleep-related or another mimic. In specialist epilepsy centres, prolonged video-EEG is also central to some presurgical assessments.
A seizure can still be missed
Scalp EEG does not record every seizure equally well. Deep, small or poorly oriented cortical generators may produce weak or absent scalp changes, and movement or muscle artefact can obscure ictal activity. A clinically convincing focal seizure is therefore not automatically disproved because the scalp EEG during an event appears inconclusive.
EEG is not a screening test for every blackout
NICE advises against routinely using EEG for transient loss of consciousness when epilepsy is not clinically suspected. Ordering EEG indiscriminately can create false reassurance from a normal test or false diagnostic certainty from nonspecific or incidental findings.
What EEG can and cannot answer
EEG can support diagnosis, classify epilepsy, identify some syndrome-specific patterns, help estimate recurrence risk after a first unprovoked seizure and capture events during monitoring. It cannot independently tell the whole story, cannot exclude epilepsy when normal, and cannot replace careful clinical history.
Related pages
First Seizure: What Happens Next? · Diagnosis and Tests · Syncope or Seizure? · Functional/Dissociative Seizures and Epilepsy
Sources
NICE NG217 — Diagnosis and assessment of epilepsy
Bouma et al. — Diagnostic accuracy of routine EEG after a first unprovoked seizure
Baldin et al. — Yield of epileptiform EEG abnormalities in incident unprovoked seizures
Kim et al., 2025 — Routine EEG yield in adults with newly diagnosed epilepsy
Epilepsy Diagnosis When Routine Ancillary Tests Are Normal, 2025
This page provides general educational information and is not individual medical advice. EEG results should be interpreted by an appropriately trained clinician in the context of the events being investigated.
Evidence checked: September 2026.