Altered neural oscillations and connectivity in the beta band underlie detail-oriented visual processing in autism

Abstract

Sensory and perceptual anomalies may have a major impact on basic cognitive and social skills in humans. Autism Spectrum Disorder (ASD) represents a special perspective to explore this relationship, being characterized by both these features. The present study employed electroencephalography (EEG) to test whether detail-oriented visual perception, a recognized hallmark of ASD, is associated with altered neural oscillations and functional connectivity in the beta frequency band, considering its role in feedback and top-down reentrant signalling in the typical population. Using a visual crowding task, where participants had to discriminate a peripheral target letter surrounded by flankers at different distances, we found that detail-oriented processing in children with ASD, as compared to typically developing peers, could be attributed to anomalous oscillatory activity in the beta band (15-30 Hz), while no differences emerged in the alpha band (8-12 Hz). Altered beta oscillatory response reflected in turn atypical functional connectivity between occipital areas, where the initial stimulus analysis is accomplished, and infero-temporal regions, where objects identity is extracted. Such atypical beta connectivity predicted both ASD symptomatology and their detail-oriented processing. Overall, these results might be explained by an altered feedback connectivity within the visual system, with potential cascade effects in visual scene parsing and higher order functions.

Keywords: Beta oscillations; EEG connectivity; Hyperconnectivity; Local perception; Vision; Visual system.

Fig. 1

Schematic representation of the task procedure. (A) Example of a trial of the visual crowding task employed in the present study, where participants with ASD and TD peers were asked to discriminate the orientation of the peripheral target letter among the four possible alternatives. There were three task conditions, labelled as strong and mid crowding when flankers appear nearby the target at a closer or farther distance, respectively, and a baseline condition where the target was displayed in isolation. (B) Three examples of stimuli employed in each trial; fillers were used to ensure a constant visual stimulation across the different task conditions. (C) Behavioural results (accuracy rates) as a function of the task condition (error bars represent SEM; * = p < .05).

Fig. 2

Event-related potential (ERPs) results. (A, B) Target evoked ERPs as a function of task conditions in the ASD and TD groups; data were obtained by averaging the activity in the posterior electrodes of interest (see upper-left insets). Activity in the grey shaded window in the time period 150–250 ms (N1) relative to the target onset was found to be differently modulated in the two groups as a function of task conditions. (C) Mean amplitude of this N1 ERP component as a function of group and task condition, revealing that while visual crowding modulated the amplitude of this component in the TD group, this modulation was absent in the ASD group (error bars represent SEM; * = p < .05).

Fig. 3

Event-related oscillatory activity reflecting detailed-oriented perception during visual crowding in ASD. (A) Scalp maps displaying the oscillatory amplitude (i.e. power) in the beta band (15–30 Hz) separately in the ASD and TD groups, together with the p-values map on the right side showing the significant cluster-corrected differences between groups (*=cluster-corrected p < .05). This analysis was performed after subtracting the oscillatory power in the baseline task condition (no flankers displayed), in order to highlight effects that emerged specifically within a crowding regime (strong and mid). (B) Time-frequency plot of the event-related oscillatory power averaged across all significant electrodes, showing a sustained decrement of beta power in the TD group after the target onset, which was not evident in the ASD group (that actually showed the opposite trend). (C) A similar analysis performed in the alpha band (8–12 Hz) did not reveal any significant differences between groups.

Fig. 4

Cortical sources of visual crowding in TD and ASD groups. (A) Source estimates (z-scored values) of neural activity for each task condition (strong crowding, mid crowding and baseline) in the time-window corresponding to the N1 (250–350 ms) ERP. In the TD group, the pattern of activation recruits an increasing bilateral portion of the occipital, temporal and (superior) parietal cortex. Such modulation in the ASD group seems to be absent, especially when contrasting the mid and strong crowding levels. (B) Contrast maps between ASD and TD obtained after subtracting the neural response in the baseline condition. The cortical regions differently activated between the two groups spanned across the frontal and anterior-temporal portions of the left hemisphere, and the right occipital lobe (as listed in detail in Table 2: minimum size of 5 vertices, p < .05 uncorrected). In all these areas, the neural activation was significantly reduced in ASD as compared to the TD group.

Fig. 5

Functional connectivity in ASD vs. TD underlying detail-oriented perception during visual crowding. Phase locking values (PLV) were estimated for the 24 sub-ROIs identified/selected through the source reconstruction (Fig. 4 and Table 2). Based on the time–frequency results, the connectivity analysis was conducted in the beta range (15–30 Hz). The connectivity matrix reports for each pair of seeds the value and direction of the statistical difference in connectivity between ASD and TD groups under a crowding regime. As assessed with “regional” cluster-corrected permutation tests, subjects with ASD showed an increased beta connectivity between the right occipital cortex and the left anterior-temporal lobe (p = .004; dashed line). The boundaries between hemispheres and anatomical regions (from left to right: orbitofrontal, prefrontal, temporal and occipital) are denoted by a solid line and different colour shades.

Fig. 6

Correlations between symptomatology in the ASD group, beta-band connectivity and behavioural performance during visual crowding. (A) Scatter plot showing the relationship between individual accuracy rate in the strong crowding condition and ASD symptomatology as measured by the Social Communication Questionnaire (SCQ) score (Current version) (a significant correlation emerged also for the mid crowding condition; see Results). (B) Scatter plot showing the relationship between individual occipital-temporal beta-band functional connectivity indexes (average PLV) and ASD symptomatology as measured by SCQ score. (C) Scatter plot showing the relationship between individual occipital-temporal beta-band connectivity and behavioural performance (i.e. accuracy rate) in the strong crowding condition. For the correlation analysis in (B), the effect of chronological age has been controlled for, whereas for the correlation analyses in (A) and (C) both the effect of chronological age and performance in the baseline task condition have been controlled for.