Resting-State Gamma Power in Young Children with ASD Participating in a Treatment Program

Background: There is growing support for the hypothesis that ASD may reflect altered structural and functional connectivity across large-scale cortical and subcortical networks (see Geschwind & Levitt 2007 for review). Resting-state EEG oscillatory activity, specifically high-frequency activity (gamma: 30-50 Hz), may reflect the aberrant functional connectivity and neural asynchrony contributing to the ASD phenotype and, hence, could serve as an important biomarker of the functional disconnection that likely occurs early during development in children with ASD. A recent study of infants at high-risk for ASD showed that frontal gamma power was reduced in high-risk infants at age 6 months as compared to age-matched, low-risk infants (Bosl et al. 2011). Another relevant study examined the relationship between gamma power and language function in typically developing (TD) children and found that individual differences in frontal resting gamma power were highly correlated with concurrent language and cognitive skills at ages 16, 24, and 36 months (Benasich et al. 2008). 

Objectives: Here, we used resting-state EEG to examine high-frequency oscillatory activity in preschool-aged children with ASD and TD, age-matched controls.  Children with ASD were enrolled in a 3-month intensive day treatment program, and EEG data was gathered prior to and at the end of treatment.  Our objectives were to (a) compare gamma activity in ASD and TD children, (b) investigate change in gamma activity after a treatment program, and (c) examine region differences between the groups of children.

Methods: Resting-state EEG was obtained for 2 minutes while the children were sitting quietly in a dark room, watching a video of bubbles. EEG data were bandpass filtered from 1 to 50 Hz. Segments containing artifacts arising from eye or muscle movements were removed from subsequent analysis. Only subjects with a minimum of 30 seconds of artifact-free data were analyzed. Data was then transformed into the frequency domain using a Fast Fourier Transform. Frontal gamma power was compared between ASD and TD children without.  In addition, EEG data prior to treatment was compared to EEG data after treatment among the children with ASD.  Lastly, we examined regional differences in the TD, the pre-treatment ASD children, and the post-treatment ASD children.

Results: We collected adequate data from 11 typically developing children, 17 children with ASD pre-treatment, and 6 children with ASD post-treatment.  The data show that pre-treatment ASD children have increased gamma power when compared to TD children, most significantly in the left frontal region (p=.018).  Additionally, there was a trend showing a decrease in frontal gamma power from pre-treatment to post-treatment (p=.095).

Conclusions:   Our very novel data show that frontal gamma power differentiates children with ASD from TD controls.  Furthermore, it appears that frontal gamma activity decreases over the course of the treatment program.  These data support the use of resting EEG as a biomarker of diagnosis and treatment outcome. As our sample size increases we will be able to further investigate these preliminary trends and their potential implications for ASD treatment programs.