ASD, face processing, and oscillatory neural activity
Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by difficulties in the social domain, repetitive behaviors, and restricted thinking [
1]. In addition to the diagnostic markers for ASD, several studies have shown that individuals with ASD may also present difficulties in face perception and emotional expression interpretation in a variety of tasks [
2,
3]. One explanation for this phenomenon suggests that, compared to typically developed individuals (TD), individuals with ASD tend to focus more on the local details of the face instead of automatically perceiving it as one figure [
4,
5].
The ability to identify emotional expression from a face requires a widespread neural network, which includes regions such as the fusiform face area, frontal regions, and occipital cortices [
6‐
8], regions that tend to show atypical activation and connectivity patterns in individuals with ASD. Moreover, it has been suggested that the ability to process the information from facial expressions stems not only from the neural activation in these regions but also from the communication between them [
9‐
12]. One way to conceptualize the relationship between face perception and communication between brain regions relates to the “top-down” and “bottom-up” perception. During face processing, one should balance between bottom-up processing, which focuses on the local features of the face, and top-down holistic interpretation by processing the face as one whole shape [
13,
14], a balance that was suggested to be atypical in ASD [
4,
5].
Using electro-and-magneto-encephalographic (EEG and MEG, respectively), studies have shown that the unique balance between top-down and bottom-up perception is modulated by the gamma-band frequency power and synchronization, a neural oscillation with a frequency between 30 and 100 Hz [
15]. While long-range connectivity in low frequencies (alpha band, 8–13 Hz) is often correlated with a priori knowledge and top-down processing [
16], the power and connectivity in gamma represent stimulus-related sensory information, local feature processing [
17], and cognitive binding of features [
18,
19]—all relevant functions for face processing. Indeed, studies have shown that an increase in gamma power and its synchronization between- and within-regions that relate to social processing, can be detected when TD individuals process upright faces compared to inverted faces or other objects [
20‐
22]. In the autistic population, there is accumulating evidence regarding deficits in alpha, beta, and gamma bands’ power and connectivity patterns during social perception [
23‐
30]. Mainly, it has been shown that during the evaluation of social stimuli, autistic individuals tend to present under-connectivity in the low-frequency bands, and over-connectivity in the gamma band, especially in magnetoencephalography (MEG) studies [
31].
However, the existing literature regarding oscillatory connectivity in ASD during social perception suffers from both inconsistencies and gaps. First, despite extensive research on high-band frequencies during social perception in ASD, the results are not consistent across the literature. While some studies suggest that deficits in gamma activation and synchronization can serve as biomarkers for autism [
32], other studies have shown that during face processing, interhemispheric connectivity in gamma can be either increased [
33‐
35] or decreased [
36,
30]. Moreover, the results are highly dependent on the experimental sample, individuals’ age, chosen task, and the type of the analysis [
31,
37]. Second, only a handful of studies examined the phase-lag connectivity of other frequencies, such as alpha and beta-band, which relate more to feedback loops and top-down reasoning. In these frequencies, the results are usually consistent, indicating that compared to TDs, individuals with ASD show hypoconnectivity [
38,
23,
39,
40,
36] except some studies showing the opposite direction by reporting an increase in alpha connectivity during social processing [
29].
Another scientific approach for investigating the top-down/bottom-up balance in ASD focused on the correlations between low and high frequencies using cross-frequencies methods such as phase–amplitude coupling (PAC) [
41]. PAC analysis allows evaluating whether the phase of a low-frequency band (usually alpha) in one region is locked to the amplitude of high-frequency bands in another region [
42]. It has been shown that during resting-state measurements, individuals with ASD present increased short-range alpha-to-gamma PAC [
43,
44]. During social perception, PAC can shed more light on the balance between top-down and bottom-up perception in ASD by examining the relationships between low and high frequencies. Indeed, studies have shown that individuals with ASD present reduced coupling during face perception in ASD for both long- and short-range connectivity, a reduction that was correlated with clinical symptoms severity [
40,
27].
Overall, the literature presents differences in the oscillatory functional connectivity during emotional face processing in ASD, compared to TD. We suggest that these differences may underlie the reported imbalance between top-down and bottom-up facial processing in the autistic population and aim to investigate the influence of Oxytocin (OT) administration on said imbalance.
It has been shown that OT, a naturally occurring hormone that relates to social communication in mammals [
45], can influence social perception in autistic individuals. For example, it has been shown that in ASD exogenous OT can modulate face perception, attention toward facial features, and identification abilities of emotional facial expressions [
46‐
51]. Although it has been found that the endogenous levels of OT in children are lower than those of TD individuals or adults [
52], the influences of exogenous OT are not consistent and depend on the context in which stimuli were presented, individual clinical and psychological parameters, as well as genetic factors such as OT receptor variants or peripheral OT level. Moreover, it is still not clear how OT influences early perceptual stages and whether these possible effects could modulate face perception in ASD.
One possible explanation for the mechanism underlying OT effects, in the context of early perception, emerges from the possible modulation of neural excitation/inhibition (
E/
I) ratio. The
E/
I model, developed by Rubenstein and Merzenich [
53], proposes that individuals with ASD present higher-than-expected neural activations due to an imbalance between local synaptic signals of excitation and inhibition. Based mainly on animal models of ASD, Lopatina et al. [
54] suggest that endogenous and exogenous OT can modify the excitation/inhibition (
E/
I) ratio through gamma-power modulation and inhibitory GABAergic pathways. Although this theory focused on local circuits, it has been proposed that disruptions in gamma feedforward course can disrupt long-range information transference across brain areas [
41,
55]. Thus, it is possible that by modulating the
E/
I ratio, OT can enhance neural connectivity and improve social perception abilities among ASD. While several studies have shown that OT indeed influences social perception and social-related neural activity in ASD, most of these studies did not examine oscillatory activity.
The current study aimed to investigate the effects of OT on top-down/bottom-up modulation in youth with ASD during the early phases of face perception. Specifically, using MEG, we measured gamma, alpha, and beta bands connectivity between three specific a priori-defined social-related regions of interest (ROIs): the left and the right fusiform and the medial prefrontal region. These regions specialize in the perception of face and facial expressions [
56‐
58] and can be modulated by OT in both TD and ASD individuals [
49,
59,
60,
61]. However, the effects of OT in these regions during the early phases of social attention are still unknown. We hypothesized that in adolescents diagnosed with ASD, OT would modulate top-down and bottom-up processing during social perception by elevating alpha- and beta-band connectivity and reducing gamma-band connectivity between the mentioned ROIs.
This is the first time such an experimental paradigm has been implemented with MEG in youth with ASD, to the best of our knowledge. Thus, although it is preliminary, we believe that the current study can offer a more comprehensive understanding of the mechanism by which OT can modulate different levels of early perceptions during face processing among autistic individuals.