The electroencephalogram (EEG) has a pivotal role in the investigation and classification of the epilepsies. The recognition of different patterns of electrical cerebral activity and their correlation with seizure types and syndromes, as well as the localisation of abnormal foci, are important in the investigation, classification and management of the epilepsies. Pattern recognition may also be important in diagnosing a range of neurological and genetic disorders and may be useful in the evaluation of encephalopathy and coma. Technology has also improved vastly with the use of multi-channel and distance recording (telemetry), co-correlation with functional magnetic resonance imaging (fMRI) and magneto-electroencephalography (MEG), and computer technology to assist with data analysis. However, computerised or automated analysis has not yet replaced the visual and generally subjective interpretation of EEG recordings in routine clinical practice.

Despite the importance of the EEG, epilepsy remains primarily a clinical diagnosis1; importantly, the EEG should not be used to make a diagnosis of epileptic seizures or epilepsy, particularly when the history of any paroxysmal events is incomplete or unclear. Apart from an incomplete and unreliable history, reliance on the findings of the EEG is the most common reason for misdiagnosing epilepsy.2 Indeed, a false diagnosis of epilepsy may have profound and potentially irreversible consequences on education, social life and employment.2 3 Consequently, the EEG has been described as one of the most abused investigations in clinical medicine.4

Unfortunately, the belief still remains that if the EEG is normal then any paroxysmal events cannot be epileptic seizures and, conversely, if the EEG is abnormal, then the diagnosis is likely (if not certain) to be epilepsy. Specifically, spikes with or without slow waves and, to a lesser extent, sharp waves are regarded as the electrical signal or hallmark of epilepsy. At least 1% of adults …