Should Every Child with Autism Have an EEG?

TED presentation by Dr. Aditi Shankardass brings up an important question--should all children with autism (or those undergoing an assessment for autism) have an EEG.  The presenter notes that in her experience in India, up to 50% of children referred with a diagnosis of autism have a seizure disorder or some other neurodevelopmental disorder.  The TED talk is posted above (7 minutes) and here are my notes from the presentation.

• 1 in 6 children suffer from developmental disorder
• Most diagnosed solely on symptoms
• Accuracy of diagnosis improved with laboratory measures
• Technology holds hope for diagnosing brain disorder
• She worked in a Harvard lab studying mathematical models to measure brain activity
• Statistical probability mapping combines EEG and statistical science
• She is setting up this technology in India

• Case study--seven year old referred with a diagnosis of autism
• EEG study was quite helpful
• Child did not have autism but had a seizure disorder
• Antieplileptic medication resulted in dramatic improvement
• She has found a significant number of children diagnosed with autism that had seizure disorders in the Indian referral population, this may be up to 50% of cases

So what about assessment guidelines for autism in the United States and Europe?  One of the best autism guidelines I could track down came from the Scottish Intercollegiate Guidelines Network.  This free guideline pdf document can be located here.  The guideline is titled:  "Assessment, diagnosis and clinical interventions for people with autism spectrum disorders. A national clinical guideline".

The guidelines provide a comprehensive review of the evidence supporting a variety of measures in the assessment process. The recommendation from the guideline related to standard EEG assessment is summarized: "Whilst epilepsy is common in children with ASD, there is no indication for an electroencephalogram (EEG) in the absence of other clinical criteria."  This recommendation comes with a citation from a 2005 review article by Kagan-Kushnir.

This review notes that seizures occur in 20-30% of children with autism and EEG abnormalities have been found in 10 to 72% of study samples.  Subclinical EEG abnormalities have been found in 6-30% of samples.  The authors note "There is insufficient evidence to recommend for or against the use of screening EEGs in autistic patients.  Given the frequency of seizure disorders in this patient population, a high index of clinical suspicion should be maintained for subtle symptoms of seizures".

The EEG technique referenced by Dr. Shankardassappears to be quantitative EEG or qEEG, a more complex EEG analysis than that done in a clinical setting.  qEEG research is ongoing at the TRANSCEND Research laboratory at Harvard directed by Dr. Martha Herbert. Dr. Herbert and colleagues have investigated a variety of EEG research measures in autism.  Recent findings indicate that autism subjects show decreased synchrony between the right and left brain hemispheres.  The clinical implications of these findings are unclear at this point.

It is unlikely that the percentage of undiagnosed clinically significant seizure disorders in autism in developed countries is high.  Nevertheless, clinicians and parents should be alert to potential subtle and non-subtle signs of seizure disorder in those with autism.  Additionally, it is not common for a child diagnosed with epilepsy to be misdiagnosed as autistic and to be "cured" by anti-epileptic therapy in the U.S. or Europe.  Although not unheard of, this misdiagnosis is reduced with comprehensive multidisciplinary assessments.  Concurrent autism and epilepsy (and EEG abnormalities) is a more likely clinical scenario than one disorder being misdiagnosed an the other.

The clinical diagnosis of autism and differential diagnosis of epilepsy continues to be one made primarily by careful history, observation and application of diagnostic criteria. Although this approach is not perfect, it does approach the reliability and validity found for other common medical conditions. When applied in a multidisciplinary setting, this approach is unlikely to misdiagnose a child with epilepsy as having autism alone.

So I think the question of need for an EEG in the context of autism should be left to the discretion of the managing physician.  Many will need an EEG as part of the assessment and differential diagnosis.  Some may not need them and the the managing physician is best able to make this distinction.

Kagan-Kushnir T, Roberts SW, & Snead OC 3rd (2005). Screening electroencephalograms in autism spectrum disorders: evidence-based guideline. Journal of child neurology, 20 (3), 197-206 PMID: 15832609

Isler JR, Martien KM, Grieve PG, Stark RI, & Herbert MR (2010). Reduced functional connectivity in visual evoked potentials in children with autism spectrum disorder. Clinical neurophysiology : official journal of the International Federation of Clinical Neurophysiology PMID: 20605520