Alpha-to-gamma phase-amplitude coupling methods and application to autism spectrum disorder.

Abstract

Adult studies have shown that a basic property of resting-state (RS) brain activity is the coupling of posterior alpha oscillations (alpha phase) to posterior gamma oscillations (gamma amplitude). The present study examined whether this basic RS process is present in children. Given reports of abnormal parietal-occipital RS alpha in children with autism spectrum disorder (ASD), the present study examined whether RS alpha-to-gamma phase-amplitude coupling (PAC) is disrupted in ASD. Simulations presented in this study showed limitations with traditional PAC analyses. In particular, to avoid false-positive PAC findings, simulations showed the need to use a unilateral passband to filter the upper frequency band as well as the need for longer epochs of data. For the human study, eyes-closed RS magnetoencephalography data were analyzed from 25 children with ASD and 18 typically developing (TD) children with at least 60 sec of artifact-free data. Source modeling provided continuous time course data at a midline parietal-occipital source for PAC analyses. Greater alpha-to-gamma PAC was observed in ASD than TD (p<0.005). Although children with ASD had higher PAC values, in both groups gamma activity increased at the peak of the alpha oscillation. In addition, an association between alpha power and alpha-to-gamma PAC was observed in both groups, although this relationship was stronger in ASD than TD (p<0.05). Present results demonstrated that although alpha-to-gamma PAC is present in children, this basic RS process is abnormal in children with ASD. Finally, simulations and the human data highlighted the need to consider the interplay between alpha power, epoch length, and choice of signal processing methods on PAC estimates.