The use of high density EEG to investigate circuit miswiring in infants at risk for autism

Thursday, May 15, 2014: 2:20 PM
Imperial B (Marriott Marquis Atlanta)
C. A. Nelson1, A. R. Levin2, M. F. Shi3 and H. Tager-Flusberg4, (1)Boston Children's Hospital, Boston, MA, (2)Neurology, Boston Children's Hospital, Boston, MA, (3)Harvard College, Cambridge, MA, (4)Psychology, Boston University, Boston, MA
Background: Autism spectrum disorder (ASD) is a complex, highly heritable disorder that involves primary impairments in language and communication.  The disorder is heterogeneous and long-term outcomes vary considerably.  In large scale studies of infants with an older sibling with the disorder, it has been estimated that approximately 1:5 such infants will eventually be diagnosed with an ASD. 

Objectives: A number of investigators have proposed that autism reflects a “connectopathy,” in which there is an overabundance of local connections and a shortage of long distance connection.  Examining neural circuitry in living children (vs. animals), however, is challenging.  Nevertheless, in our work we have made use of advanced signal processing tools to examine the development of neural circuitry. 

Methods: As part of our ongoing longitudinal study of high risk infants, task and resting EEG was collected on 208 infants ranging in age from 3 to 36 months of age.  EEG was recorded from 128 scalp sites, using EGI sensor nets.  Three components of resting EEG were examined.   First, EEG coherence was examined as indexed by the mean coherence between all electrode pairs represented in the 10-20 system, with a particular emphasis on the delta and theta bands. Second, alpha rhythm in electrodes over the primary motor cortex was measured, which typically decreases in amplitude when one person observes another performing a motor action, a phenomenon known as mu suppression.  Third, coupling between the phase of the theta band pattern and the amplitude of the gamma band pattern was also examined.

Results: EEG coherence results has demonstrated a peak in mean coherence (particularly in the delta and theta bands) at 24 months in children who develop ASD, but at 12 months in children who do not develop ASD.  These findings demonstrate that infants with an older sibling with ASD show a corresponding peak in mean coherence in the beta and gamma bands that is not seen in children with a typically developing older sibling.  Mu suppression results have shown that at 9 to 18 months, some of the infants who develop ASD show a lack of alpha suppression in primary motor cortex (mean alpha power 20.5 μV2 at 9 months and 14.6 μV2 at 18 months in high risk infants who develop ASD, versus 6.7 μV2 at 9 months and 4.9 μV2 at 18 months in high risk infants who do not develop ASD), a phenomenon which has previously been seen in older children and adults with ASD.  Finally, coupling analyses have shown increased phase amplitude coupling (PAC) in the temporal regions in infants 3-36 months who develop ASD (Mean modulation index 0.95 left temporal, 0.82 right temporal) compared to those who do not (Mean modulation index 0.21 left temporal (p = .003), 0.08 right temporal (p = .007).

Conclusions: Preliminary EEG analyses thus far appear promising both as an endophenotype of high ASD risk and as a potential biomarker for ASD diagnosis.