Epilepsy is diagnosed through a combination of tests rather than a single definitive exam. The process typically starts with an EEG (electroencephalogram) to measure brain electrical activity, followed by brain imaging and sometimes blood work or genetic testing. No single test can confirm or rule out epilepsy on its own, which is why doctors often layer several together to build a complete picture.
EEG: The First-Line Test
An EEG is almost always the first test ordered when epilepsy is suspected. Small electrodes are placed on your scalp to record the brain’s electrical patterns. Doctors look for specific abnormal spikes and waves called interictal epileptiform discharges, which are highly specific for epilepsy at about 98%. That means if these patterns show up, there’s very little chance they’re a false alarm.
The catch is sensitivity. A standard 20 to 30 minute EEG only picks up these telltale patterns in 20 to 50 percent of people with epilepsy. In adults after a first unprovoked seizure, that number drops to just 17 percent. So a normal EEG does not mean you don’t have epilepsy. It simply means the test didn’t catch abnormal activity during that brief window.
If the first EEG comes back normal, your doctor may order a repeat or a sleep-deprived version. Sleep deprivation increases the frequency of abnormal brain activity: roughly 41 percent of patients show epileptic patterns on a sleep-deprived EEG compared to only 13 percent on a standard follow-up. This boost is especially pronounced in generalized epilepsy, where sensitivity jumps to about 64 percent after sleep deprivation, though it helps less with focal epilepsy (around 17 percent).
How to Prepare for an EEG
Preparation matters because anything on your scalp can interfere with the electrode signals. Shampoo your hair the night before, but skip conditioner, gel, hairspray, and any other styling products. Avoid caffeine, coffee, tea, chocolate, soft drinks, and alcohol for 12 hours before the test. Continue taking your regular medications unless your doctor specifically tells you otherwise.
If you’re scheduled for a sleep-deprived EEG, you’ll typically be asked to stay up all night or significantly limit sleep the night before. The goal is to increase the chance your brain produces the abnormal patterns doctors are looking for.
Prolonged and Video EEG Monitoring
When a routine EEG isn’t enough, prolonged ambulatory video-EEG monitoring offers a much wider window. You wear a portable EEG device for several days while a small camera records your movements and behavior. This allows doctors to match what’s happening in your brain with what’s happening in your body during an actual event.
Recording duration in these studies averages about 5.7 days and can run up to ten. Each additional day of recording adds diagnostic value through the eighth day. About 41 percent of all ambulatory video-EEG recordings capture a diagnostic event, and 14 percent of those events are caught between days five and eight. For classifying seizure types, the yield reaches 75 percent. This type of monitoring is particularly useful when doctors need to distinguish epileptic seizures from other types of episodes.
Brain Imaging With MRI
An MRI scan looks for structural problems in the brain that could be causing seizures: scar tissue, tumors, malformations of brain development, or damage from a prior stroke or injury. Unlike an EEG, which measures electrical activity, an MRI creates detailed images of the brain’s anatomy.
Not all MRIs are equal for epilepsy. A 3 Tesla (3T) MRI is more sensitive than the standard 1.5T machines found in many imaging centers, thanks to better spatial resolution and a stronger signal. Epilepsy-specific MRI protocols include specialized sequences that highlight subtle abnormalities ordinary brain scans might miss, particularly in structures deep in the temporal lobe where many seizures originate. If your initial MRI was done on a lower-strength machine or without an epilepsy protocol, a repeat scan at a specialized center can sometimes reveal a previously hidden cause.
Blood Tests and Prolactin Levels
Standard blood work plays a supporting role. Tests for blood sugar, electrolytes, kidney function, and liver function help rule out metabolic causes of seizures. These aren’t epilepsy tests per se, but they can identify treatable conditions that mimic it.
One specialized blood test measures prolactin, a hormone that surges after certain types of seizures. When drawn 10 to 20 minutes after a suspected event, prolactin levels that are three to four times above baseline strongly suggest a generalized tonic-clonic or complex partial seizure rather than a psychogenic (non-epileptic) episode. The American Academy of Neurology recommends this test as a useful tool for telling the two apart, though it’s only helpful when blood can be drawn within that narrow post-event window and works less reliably for other seizure types.
Genetic Testing
Genetic testing isn’t routine for every person with seizures, but it’s increasingly important in specific situations. The American Epilepsy Society identifies several profiles where genetic panels or more comprehensive genome-level testing is recommended:
- Childhood-onset epilepsy, especially severe forms beginning before age 3
- Drug-resistant epilepsy of unknown cause
- Family history of epilepsy in two or more first-degree relatives
- Unexplained epilepsy, where imaging, lab work, and clinical evaluation haven’t identified a structural, metabolic, infectious, or other acquired cause
- Additional neurologic findings such as intellectual disability, autism, movement disorders, or cerebral palsy alongside seizures
Identifying a genetic cause can change treatment in meaningful ways. Some genetic epilepsies respond better to specific medications, and others have medications that should be avoided because they can worsen seizures. For families, a genetic diagnosis also clarifies recurrence risk for future children.
Ruling Out Conditions That Mimic Epilepsy
A significant part of epilepsy testing is making sure seizures aren’t actually something else. Fainting (syncope) is one of the most common mimics, and the two can look surprisingly similar. Tilt table testing, where you’re strapped to a table that shifts from horizontal to upright while heart rate and blood pressure are monitored, can help distinguish the two.
Research from a study published in Neurology identified practical ways to tell them apart. In syncope, people experience a loss of muscle tone (going limp), which essentially never happens during a convulsive seizure. Jerking movements in syncope tend to be few (a median of 2, with a maximum of about 19) and asynchronous, while convulsive seizures produce many more rhythmic jerks (a median of 48, ranging from 20 to nearly 200). This “10/20 rule” offers a useful guideline: fewer than 10 jerks points toward fainting, more than 20 toward a seizure. Muscle stiffening and extension of the limbs is about eight times more likely in a seizure than in syncope.
Psychogenic non-epileptic seizures are another common mimic, accounting for a substantial portion of cases referred to epilepsy monitoring units. Video-EEG is the gold standard for identifying these, because the brain’s electrical activity remains normal during the episode even though the person’s body appears to be seizing.
What the Testing Process Looks Like Overall
For most people, the diagnostic path follows a predictable sequence. After a first seizure, you’ll get a neurological exam, blood work, a routine EEG, and usually an MRI. If the EEG is normal, which it often is, a repeat or sleep-deprived EEG comes next. If seizures continue but remain unclassified, prolonged video-EEG monitoring may be recommended. Genetic testing enters the picture when the cause remains unclear or when the clinical picture suggests a hereditary syndrome.
The process can feel frustratingly slow, especially when the first round of tests comes back normal. But normal initial results are common and expected given the limited sensitivity of a single EEG. Each additional test narrows the possibilities, and the combination of electrical, structural, and sometimes genetic data is what ultimately leads to a diagnosis.