The Effect of Visual Stimulation on Post-Stimulus Alpha Oscillations in Autism Spectrum Disorder

It is well known that neural and behavioral modulations by sensory stimuli are affected by the preceding pattern of neural oscillations, especially that in the alpha range. However, it is unknown how alpha oscillations in the post-stimulus interval are themselves affected by the sensory features of preceding visual stimuli. Since alpha activity has been shown to reflect an inactive state of the cortex, stronger alpha activity after strong visual stimulation may reflect ‘on-demand’ inhibitory regulation of neural excitation in visual areas. This inhibitory regulation has been shown to be atypical in people with autism spectrum disorder (ASD) (Keehn et al., 2017). Abnormal neural inhibition and top-down connectivity, as demonstrated in ASD (Cornew et al., 2012; Kessler et al., 2016), may affect the post-stimulus alpha modulation in this population. In the current study, we investigated how the intensity of visual input affects magnetoencephalography (MEG) oscillations in the ‘post-stimulus’ interval in people with and without ASD. We found that in both experimental groups stronger visual stimulation was associated with a greater post-stimulus increase in the high-alpha/low-beta power (alpha-beta ‘rebound’) and an increase of the individual alpha peak frequency over posterior MEG sensors. We localized the neural source of the effect of the preceding visual stimulation on alpha/beta power using two different methods. First, we localized the sources of the MEG sensor-based effects using standard approaches that include the total MEG signals (sLORETA). Complementing this approach, we used a Common Spatial Pattern (CSP) transformation of the sensor space data (Blankertz et al., 2008), followed by source localization of the CSP patterns. CSP transformation allows for the maximization of the power difference between experimental conditions. Common spatial patterns with the largest between-condition power difference are expected to have a more focused source in the target area relevant to the task (Herrojo Ruiz et al., 2017). Using sLORETA we found that the effect of visual stimulation in all participant groups is broadly distributed over posterior cortical areas involved in visual processing (Pascual-Marqui, 2002). The CSP analysis, using six patterns associated with maximal between-condition power differences, did reveal very localized sources, as expected, however there was a large degree of inter-subject variability in the spatial location. Accordingly, on the group level, CSP-based source localization failed to recover a coherent spatial pattern of the stimulation effect. Our preliminary results based on normalized power differences between the stimulation conditions did not find that the group factor (ASD vs NT) had an effect on the post-stimulation alpha-beta activity. We interpret the lack of significant differences to suggest that post-stimulus inhibition (indexed by alpha activity) is preserved in adults with ASD and normal IQ. At the same time, it is still possible, that this inhibition is impaired in low-functional individuals with ASD. It remains unclear whether the effect of a post-stimulus alpha-beta increase we found reflects local processes in brain areas involved in processing of visual information or the top-down inhibitory regulation by the higher-order cortical areas. To identify the nature of the stimulus-related enhancement of the alpha-beta rebound we plan to perform directed connectivity analysis (e.g. using Granger Causality or Phase slope index) to see the effect of preceding stimulation on the direction of information transfer in the alpha-beta range.