Hypsarrhythmia: the EEG pattern of infantile spasms

Short answer: hypsarrhythmia is the classic EEG pattern of infantile spasms. Between spasms, the brain waves are very high in voltage, disorganized, and full of spikes coming from many areas at once.^1,2 It is one part of the triad that defines West syndrome, and clearing it on a repeat EEG is part of how doctors judge whether treatment is working.^3,4

An EEG (electroencephalogram) records the brain's electrical activity through small sensors on the scalp. For a baby with suspected spasms, experts recommend a prolonged video-EEG recorded awake and asleep.⁵ The US consensus calls for early recognition, and for a timely EEG to judge the response once treatment starts.⁶

What is hypsarrhythmia?

A 2015 study described it as the "classic interictal electroencephalographic pattern associated with infantile spasms," marked by high voltage, disorganization and multifocal independent epileptiform discharges.¹ "Interictal" means between spasms, not during them.

A 2020 review describes the pattern as a chaotic mix of very high-amplitude slow waves, with spikes and sharp waves that vary in amplitude, shape, duration and location.²

What hypsarrhythmia looks like on EEG

Described in words, a hypsarrhythmia EEG has four main features:^1,2

• Very high voltage. The slow background waves are very large.
• Disorganization. The usual orderly rhythms of a baby's brain are lost.
• Spikes from many places. Epileptiform discharges (spikes and sharp waves) appear independently in many areas, not from one spot.
• Constant change. The size, shape and location of the spikes keep changing.

During a spasm itself, the EEG often looks different. Several patterns are seen. The most common is a generalized slow wave followed by a drop (attenuation) in background activity in all regions.⁷

Hypsarrhythmia vs normal EEG

Compared with a normal recording, hypsarrhythmia stands out for its high voltage, disorganization and spikes from many places.¹

Telling an abnormal EEG from a normal one is usually easy. In a study where six blinded pediatric EEG experts rated 22 five-minute awake EEG samples, they agreed very well on whether each EEG was abnormal (kappa 0.89, where 1.0 is perfect agreement).¹ Deciding whether an abnormal EEG counts as hypsarrhythmia is much harder (see below).

Modified hypsarrhythmia

Long recordings show that hypsarrhythmia is "highly variable and dynamic." The variations, together called modified hypsarrhythmia, include:⁸

• More synchronized activity between the two sides of the brain
• An asymmetric (lopsided) pattern
• A consistent focus of abnormal discharges
• Episodes where the activity suddenly drops off (attenuation)
• Mostly high-voltage slow waves with few spikes

Other reviews also list focal discharges and a burst-suppression-like pattern.²

Why the EEG needs to include sleep

The pattern changes with the baby's state. The ILAE says the abnormality is enhanced in drowsiness, in sleep and on waking, and early on it may only be seen in sleep or on waking.⁹ Long recordings show it can also fade or disappear in REM (dreaming) sleep, and look more normal briefly on arousal and during a cluster of spasms.⁸ What the EEG shows depends on how long it records, whether the baby is awake or asleep, and the structure of the baby's brain.⁸ Either way, a short recording can miss or misjudge the pattern, and the EEG should include sleep.⁵ In the trial behind the ACTH label, response was judged on a full sleep-cycle video EEG.¹⁰

Why experts sometimes disagree

Hypsarrhythmia sounds simple to spot, but the research says otherwise. In the same six-rater study, agreement on whether hypsarrhythmia was present was poor (kappa 0.40), and agreement on modified hypsarrhythmia was 0.47. Agreement on "disorganization" was only 0.22. The raters still felt confident in their calls.¹

An international expert group could not agree on a definition of hypsarrhythmia.⁴ To make readings more consistent, researchers built a simpler score, the BASED score (Burden of Amplitudes and Epileptiform Discharges). In 40 EEG clips, agreement on hypsarrhythmia was 0.87 with the BASED score against 0.09 with traditional reading.¹¹ A 2026 Danish study found the same direction: agreement on an epileptic encephalopathy (Gwet's AC1) was 0.88 with BASED against 0.62 with traditional scoring.¹²

What this means for families. Many babies with infantile spasms never show classic hypsarrhythmia. One expert review warns that relying on it too much could lead to standard treatment being wrongly withheld.¹³ If your baby has spasms and the EEG is abnormal but "not hypsarrhythmia," ask whether treatment should still start now.

Role in diagnosis

The ILAE defines infantile epileptic spasms syndrome (IESS) by epileptic spasms starting between 1 and 24 months of age. West syndrome is the subset with the full triad: spasms, hypsarrhythmia and developmental plateauing or regression.³ In a 2026 review of 28 guidelines from around the world, 27 recommended an EEG.¹⁴

Role in judging response to treatment

Hypsarrhythmia is also used to judge whether treatment has worked:

• Consensus definition. The West Delphi group defined the main electroclinical outcome as no more spasms plus resolution of hypsarrhythmia.⁴
• US cohort studies. In a 23-center US study, a sustained response meant the last spasm on or before day 14, no hypsarrhythmia on an EEG 2 to 4 weeks after starting treatment, and remission lasting to day 30. Babies without hypsarrhythmia before treatment were more likely to respond early.¹⁵ In a related study, a first treatment counted as failed if spasms continued or hypsarrhythmia persisted.¹⁶
• Drug label. For ACTH, the label counted a baby as a responder only with complete suppression of both spasms and hypsarrhythmia on a full sleep-cycle video EEG 2 weeks after starting.¹⁰

Because agreement between readers is imperfect, the same inconsistency can affect these response decisions and research results.¹

Hypsarrhythmia treatment

Hypsarrhythmia is not treated on its own. It is treated by treating the infantile spasms, usually with ACTH, oral prednisolone or vigabatrin. Vigabatrin is the first choice when the cause is tuberous sclerosis complex.¹⁴ Response is usually checked at about 14 days.¹⁴ A short duration of hypsarrhythmia is one of the outcome factors that care can change.¹⁷

If vigabatrin is chosen, accurate doses matter from the first day. Every vigabatrin form is the same medicine.¹⁸ The ready-to-use liquid (labeled for infantile spasms from 1 month to 2 years, like the powder) needs no mixing; in one small study funded by its maker, doses prepared from it were more accurate than doses mixed from powder.^18,19

Why an urgent EEG matters

Time to treatment is linked to how children do later. In the UK Infantile Spasms Study, each step up in the delay before effective treatment was linked to a lower development score at age 4.²⁰ In the ICISS trial, a longer delay was linked to lower development scores at 18 months.²¹ The US consensus calls for early recognition, and a timely EEG to judge the response to treatment.⁶

Delay is common: in a survey of 100 families, the median time from the first spasm to a doctor who recognized it and started treatment was 24.5 days.²² In a review, treatment within 4 weeks of onset made a good developmental outcome about 1.5 times as likely as later treatment.²³ If you see possible spasms, film them, and ask your doctor promptly about an EEG; spasms are one of several kinds of seizures in babies.