Patterns of Epileptiform Activity in Autism as Revealed by MEG

Background: It is well established that children with an autism spectrum disorder (ASD) are at higher risk for developing clinical seizures than the general population, with 25-30% reported to have epilepsy by the time they reach adolescence. Interestingly, many children with autism who do not have clinical seizures are nevertheless reported to have epileptiform EEG examinations. Current clinical guidelines for the evaluation of children with autism indicate that EEG is warranted whenever clinical seizures are suspected or there is a regression in language skills. We have previously demonstrated that EEG reveals epileptiform activity in >45% of sedated autistic children with a history of regression, whereas simultaneous MEG reveals epileptiform activity in >60%, with the increased sensitivity for MEG mostly reflecting MEG’s superior ability to detect activity within the upper bank of the superior temporal gyrus.

Objectives:  The current study sought to better define MEG and EEG patterns of epileptiform activity in autism, especially with respect to other clinical factors including gender, presence of clinical seizures, intelligence, level of language function, and regressive versus developmental delay language profiles.

Methods: Forty minutes of whole-head MEG and simultaneous EEG data were collected from more than 80 sedated children (ages 3-16) diagnosed with autistic disorder (DSM-IV, ADI-R, ADOS).  MEG data were collected using a 306 channel biomagnetometer system equipped with both planar gradiometers and magnetometers. EEG data were acquired with 19-60 electrodes configured in an augmented 10-20 array. Sedation was achieved with either oral clonidine and chloral hydrate, or IV propofol, as needed. The MEG and EEG data were analyzed for epileptiform transients with multiple dipole modeling used to characterize the origins of individual discharges.

Results: Although less that 15% of the children had a history of clinical seizures, MEG revealed epileptiform activity in >60%. Epileptiform abnormalities included isolated spikes and sharpwaves, and also sharp burst discharges. Perisylvian activity was more commonly seen than activity in other brain regions, but most children showed a multifocal, bihemispheric profile of epileptiform abnormalities. For approximately 20% of the children with an abnormal MEG, the simultaneous EEG was within normal limits. This pattern was most commonly seen for children with MEG abnormalities restricted to the superior temporal plane. Contrary to initial expectations, epileptiform activity was NOT more prevalent in children with history of language regression than those with the more typical pattern of developmental delay (early-onset) autism. Low functioning children were twice as likely to show epileptiform activity as high functioning children, but the severity of autism per se, was not a significant factor. Children that were verbal but without functional language were more likely to show epileptiform abnormalities than children that were high functioning or children that were completely non-verbal.

Conclusions:   MEG reveals epileptiform abnormalities in the majority of children with autistic disorder. This activity is most prominent in low functioning children with non-functional language, with the high incidence of peri-sylvian activity suggesting that the epileptiform activity is a direct neurobiological cause or consequence of disordered language circuitry.