EEG Variability at 3 Months Correlates with Autism Outcomes

Thursday, May 11, 2017: 10:50 AM
Yerba Buena 9 (Marriott Marquis Hotel)
A. R. Levin1, H. M. O'Leary1, A. S. Mendez Leal2, A. Acosta3, K. J. Varcin4, J. M. Mayor Torres5, H. Tager-Flusberg6 and C. A. Nelson7, (1)Neurology, Boston Children's Hospital, Boston, MA, (2)Harvard College, Cambridge, MA, (3)Harvard, Cambridge, MA, (4)Telethon Kids Institute, Perth, Australia, (5)Harvard Medical School, Cambridge, MA, (6)Psychological and Brain Sciences, Boston University, Boston, MA, (7)Boston Children's Hospital, Boston, MA

There are now a plethora of studies using EEG to elucidate neural connectivity in autism. While within-subject variability in EEG responses to repetitive stimuli are often thought to be a barrier to be overcome in such studies, recent work suggests that variability itself may be an important marker of underlying neurobiological processes, rather than a barrier to be overcome. To this end, elevated intra-participant variability in response to repetitive trials has now been demonstrated in multiple studies of children with autism spectrum disorder (ASD).1

Increasing evidence also suggests that alterations in brain development during infancy precede the manifestation of overt, behavioral signs of autism spectrum disorder (ASD). A key finding in infant studies, however, is that findings present in children with ASD may be different from findings in infancy that portend a later ASD diagnosis. For example, prior studies from our group have shown that high frequency frontal power is decreased in 3-month-old infants at high risk for ASD, but increases excessively over the first 3 years of life in infants who later develop ASD.

Here, we therefore assess the extent to which trial-by-trial within-participant variability, previously shown to be elevated in children and adults with ASD, is altered in infants who are later diagnosed with ASD.

Objectives: We analyzed response variability in 3-month-old infants later diagnosed with ASD, compared to infants who were later found not to have ASD.

Methods: ERP data were acquired on 3-month-old infant siblings of children with ASD (high risk; HRA; n = 41) and 3-month-old low risk controls (LRC; n = 16) as part of a prospective, longitudinal investigation. For each infant, we used inter-trial alpha phase coherence (ITPC) to measure variability of the P150 response across trials to a standard native consonant-vowel auditory stimulus (/da/). Diagnosis of ASD(+) or lack thereof(-) was determined at 24-36 months using the Autism Diagnostic Observation Schedule (ADOS) and best clinical estimate (HRA+: n = 11; HRA-: n = 14; LRC-: n = 16).

Results: 3-month-olds later diagnosed with ASD had significantly increased P150 response variability (p < .05), compared to 3-month-olds later found not to have ASD.

Conclusions: Increased intra-individual response variability to a repetitive stimulus is present in 3-month-olds who will later develop ASD. This suggests that excessive variability is present prior to the onset of behavioral manifestations of ASD, and thus may be an early marker of altered brain function on the pathway to ASD symptoms.


1. David N, Schneider TR, Peiker I, Al-Jawahiri R, Engel AK, Milne E. Variability of cortical oscillation patterns: a possible endophenotype in autism spectrum disorders? Neurosci Biobehav Rev. 2016.