Analysis of Alpha Power in 9-Month Old Infants at High-Risk of ASD

Background:  Previous studies found differences in frontal alpha waves characterizing infants at high-risk (HR) for ASD when compared to low-risk (LR) age-matched groups. In particular, decreased frontal alpha power and altered hemispheric asymmetry have been described during the first year of life (Tierney et al., 2012; Gabard-Durnam et al., 2015).  Interestingly, the developmental trajectories of alpha asymmetry proceeded in opposite directions between groups:  whereas LR infants showed an initial relative right frontal power at 6 months, HR infants presented left frontal asymmetry initially, decreasing with age at 12 and 18 months (Gabard-Durnam et al., 2015). 

Objectives:  

Given the differences described for alpha bandwidth between HR and LR infants, we aimed to examine alpha power at 9 months, around the time when alpha asymmetry shift is predicted to take place. We thus asked whether alpha power differed between HR and LR groups in frontal and posterior regions.

Methods:  

Resting EEG data at 9 months of age were collected from HR infants (n=25; F: 37.5%; M: 62.5%; including 4 ultra high-risk subjects, defined as having more than one sibling with ASD), and LR infants (n=16; F/M: 50%) using high-density 128-hydrogel nets (EGI inc). In NetStation, data were filtered at 1–50 Hz, divided into 1 sec. segments and re-referenced using an average reference.  Segments containing artifacts were visually inspected and removed from the analysis. Regions of interest (ROIs) were defined as described before (McEvoy et al. 2015). We calculated absolute alpha power at 6-9 Hz, a well-defined band for infant alpha EEG (Stroganova et al., 1999; Marshall et al., 2002).

Results:  

Absolute alpha power was analyzed at frontal and posterior locations. Independent samples t-test revealed a slight trend towards decreased frontal alpha in HR compared to LR (mean ± SEM: 11.14 ± 1.3 vs.12.96 ± 2.0, respectively; p=0.43, t=0.80), with no significant differences at posterior sites (HR: 14.12 ± 1.4, LR:13.60 ± 1.9; p=0.72, t=0.36). Although frontal hemispheric asymmetry scores (right-left) were similar (HR: -0.25 ± 0.73; LR: -0.75 ± 1.6; p=0.92, t=0.11); posterior-frontal asymmetry scores showed highly significant differences (HR: 2.97 ± 0.56; LR: 0.65 ± 0.66; p=0.013, t=2.60). In spite of being limited by our sample size, UHR infants generally followed the trend of their assigned group (HR).

Conclusions:

Overall, our data suggest that alpha power is a useful indicator of ASD risk in infancy. The significant reduction in posterior-frontal scores, together with the trend towards reduced frontal alpha, indicate that relevant rearrangements in the distribution of alpha occur at this critical developmental age.