Dysregulated Local Oscillatory Connectivity of the Visual System in Autism Spectrum Disorder

Background:

Individuals diagnosed with Autism Spectrum Disorder (ASD) frequently report sensory overload (both hypo- and hyper-sensitivities), and display a distinct sensory processing style. However, the neurophysiological principles underlying this remain largely unknown. Recent non-invasive magnetoencephalography (MEG) autism research suggests one candidate mechanism may be disorganised local oscillations (neuronal rhythms) produced in response to sensory stimuli, combined with reduced top-down modulation.

Objectives:

We set out to investigate the oscillation-based connectivity within the visual system of ASD participants and matched controls. Specifically, we quantified local dysregulation through low-frequency alpha (7-13Hz) to high-frequency gamma (40-100Hz) phase-amplitude coupling (PAC); and feedforward versus feedback connectivity, using non-parametric granger causality.

Methods:

Whilst undergoing MEG, 17 participants (mean age = 16.63) diagnosed with ASD and 17 matched controls (mean age = 17.43, 15 analysed at present) performed an interactive visual paradigm, with embedded visual grating, designed to elicit non phase-locked high-frequency (40-70Hz) gamma oscillations. Both groups had movement less than 5mm and similar numbers of trials (mean ASD = 62.46, mean control = 62.33). All participants completed the Autism Quotient and Glasgow Sensory questionnaires, to asses the severity of ASD sensory symptoms.

Results:

As expected, presentation of the visual grating was accompanied by increases in gamma-band (40-70Hz) power and decreases in alpha-band (8-13Hz) power within the ventral occipital cortex. There were no statistical differences in power between groups at either frequency. However, there were differences between groups in the coupling (PAC) between the amplitude of gamma-band oscillations and the phase of alpha-band oscillations extracted from virtual electrode time courses in primary visual cortex (V1). The ASD group showed PAC between higher alpha (12-13Hz) and gamma frequencies (60-80Hz) than the control group (9-10Hz and 50-70Hz, respectively). We argue that this could be indicative of dysregulated oscillatory activity and an excitation-inhibition imbalance in autism. Next, virtual electrode time-courses were extracted from areas V1 and V4 to quantify directed functional connectivity from V1 to V4 (feedforward) and V4 to V1 (feedback), using Granger causality. Whilst feedforward connectivity was equivalent in both groups and elevated at gamma frequency (40-80Hz) in particular, V4 to V1 feedback connectivity was significantly reduced in the ASD group, especially at alpha frequency (~10Hz).

Interestingly, the strength of alpha-gamma PAC (9-10Hz; 50-70Hz) was correlated with the ratio between feedforward and feedback connectivity, for the control but not ASD group. Neurophysiologically, this suggests that processing within the visual system is locally segregated in autism, resulting in a stronger dissociation between peak gamma frequencies for PAC (~75Hz), power (~55Hz) and forward connectivity (~45Hz).

Finally, these MEG data will be regressed against self-report questionnaire data, to determine whether local connectivity is related to ASD sensory symptoms.

Conclusions:

Overall, our work suggests that the complex interplay of stimulus-related alpha and gamma oscillations within the human visual system could be dysregulated in autism. The ASD group showed typical patterns of oscillatory power and feedforward connectivity, but dysregulated oscillatory coupling between frequency bands (PAC) and reduced feedback alpha-band connectivity. These findings have implications for emerging neurocognitive theories of sensory processing in autism.