EEG Analysis of Visual Attention Under Perception of Affective Prosody in Children with Autism Spectrum Disorder

Multiple object tracking (MOT) tasks are widely used to examine attention impairments in autism spectrum disorder (ASD); however, existing research has predominantly relied on behavioral measures and has barely incorporated physiological data, such as electroencephalography (EEG). Additionally, how cognition and emotion interact in individuals with ASD during MOT tasks remains unclear.

In this study we addressed these gaps using a publicly available dataset containing EEG recordings and clinical measurement from 28 children with ASD and 28 typically developing (TD) individuals. Participants listened to speech recordings reflecting prosodic happiness and sadness while completing three attention-level tasks: neutral image viewing (low-attention), one-target four-disc MOT-4 (intermediate-attention), and one-target eight-disc MOT-8 (high-attention). For each participant's EEG data, we calculated the theta-beta ratio (TBR) for each channel and then extracted five EEG features: averaged TBR for all channels (allTBR), averaged TBR for channels in the frontal brain region (fTBR), averaged TBR for channels in the temporal brain region (tTBR), averaged TBR for channels in the central brain region (cTBR), and averaged TBR for channels in the parietal brain region (pTBR). Correlation analysis was also conducted to explore the association between EEG features and clinical measures.

Findings showed that: (1) the interaction effect between attentional level and emotion was significant for allTBR and tTBR; (2) Significant main effects of group were observed for all five TBR metrics (allTBR, fTBR, tTBR, cTBR, and pTBR), with the ASD group showing higher values than the TD group; (3) only tTBR showed a significant difference between happiness and sadness stimuli during the MOT-8 task after Bonferroni correction; (4) during the MOT-4 task, the TD group had higher correlations between five EEG features and clinical measures than the ASD group; and (5) the tTBR finding during the MOT-8 task suggests that emotional valence may modulate neural processing in the temporal region under high cognitive load.

These findings could provide new insights into attentional impairments and cognition-emotion interactions in ASD, offering potential implications for the development of clinical intervention strategies.