Functional Connectivity Analysis of Children with Autism under Emotional Stimulation

Background Introduction

Autism Spectrum Disorder (ASD) is a complex neurodevelopmental disorder, primarily characterized by deficits in social interaction and communication abilities, as well as repetitive behaviors and restricted interests. One of the core features of ASD is emotional processing deficits, which directly affect patients’ social skills and quality of life. Although ASD research has been ongoing for many years, its neural mechanisms are still not fully understood, especially regarding brain functional connectivity patterns during emotional processing. Functional brain connectivity analysis is an important method to study the neural mechanisms of ASD, and Electroencephalography (EEG), as a non-invasive technique that can record real-time brain electrical activity, is a powerful tool for studying brain functional connectivity.

However, most existing EEG studies focus on spontaneous brain activity and rarely involve brain functional connectivity under emotional stimulation. Therefore, exploring the differences in brain functional connectivity in children with ASD under emotional stimuli not only helps to understand the neural mechanisms of ASD but also may provide potential biomarkers for early screening and intervention of ASD.

Source of the Paper

This paper was co-authored by Chang Cai, Jiahui Wang, Jun Lin, Kang Yang, Huicong Kang, Jingying Chen, and Wei Wu, who are respectively from the National Engineering Research Center at Central China Normal University, the Department of Neurology at Tongji Hospital, and Songjiang Hospital of Shanghai Jiao Tong University School of Medicine. The paper was published in August 2021 in the Journal of LaTeX Class Files, titled “Functional Connectivity Analysis of Children with Autism under Emotional Clips.”

Research Process

Study Subjects and Experimental Design

The study recruited a total of 64 children, including 32 children with ASD and 32 typically developing (TD) children. The inclusion criteria for ASD children included meeting DSM-V diagnostic standards, being aged between 2 and 7 years, and having no severe respiratory diseases or epilepsy and other brain-related conditions. TD children were recruited from local kindergartens, matched in age with the ASD group, and had no history of mental disorders or developmental delays.

The experiment used the portable EEG device Emotiv Epoc, which is equipped with 14 EEG channels capable of recording electrical activity in different areas of the brain. The experiment was conducted in a controlled virtual environment where children watched two 60-second video clips: one inducing positive emotions and the other negative emotions. EEG data were recorded while the videos were being watched.

Data Preprocessing and Analysis

EEG data preprocessing included three steps: first, using Common Average Reference (CAR) to eliminate DC offset; second, using the EEGLAB toolbox to remove artifacts caused by eye movements and muscle activity; finally, filtering using a linear bandpass filter (1 Hz to 45 Hz) to remove noise and high-frequency artifacts.

After preprocessing, researchers extracted five frequency bands from the EEG data: θ wave (4-8 Hz), α wave (8-12 Hz), low β wave (12-16 Hz), high β wave (16-25 Hz), and γ wave (25-45 Hz). Then, four functional connectivity metrics were used to analyze brain functional connectivity: Coherence, Phase-Locking Value (PLV), Phase Lag Index (PLI), and Weighted Phase Lag Index (WPLI).

Brain Functional Connectivity Difference Analysis

To explore the differences in brain functional connectivity between children with ASD and TD under emotional stimuli, researchers calculated the average connectivity values of the four connectivity metrics across five frequency bands and compared the differences between the two groups. The results showed that under negative emotional stimuli, children with ASD exhibited poorer mutual correlation and phase synchronization in the low-frequency range, while showing stronger intracerebral network coordination in the high-frequency range. Under positive emotional stimuli, TD children showed stronger functional connectivity between brain regions, whereas ASD children exhibited higher phase synchrony among brain regions.

Additionally, the study found that in the θ frequency band, ASD children exhibited more long-range connections, particularly between the left and right hemispheres. In the α frequency band, ASD children showed fewer long-range connections, indicating a preference for local detail processing. Under positive emotional stimuli, the functional connectivity of the frontal lobe in ASD children was enhanced, while under negative emotional stimuli, the temporal lobe showed increased activity.

Classification and Feature Selection

To verify the potential of brain functional connectivity features in ASD screening, researchers used Support Vector Machine (SVM) and Minimum Redundancy Maximum Relevance (MRMR) algorithms for classification. Through MRMR feature selection, researchers selected the most representative subset from all features and used SVM for classification. The results showed that under positive emotional stimuli, using 15 features achieved a classification accuracy of 85%, while under negative emotional stimuli, using 20 features achieved a classification accuracy of 87%.

Research Results and Conclusions

The main findings of the study include: 1. Impact of Emotional Stimulation on Brain Functional Connectivity: Children with ASD exhibited lower low-frequency connectivity under negative emotional stimuli, while showing stronger intracerebral network coordination in the high-frequency range. Under positive emotional stimuli, ASD children showed higher phase synchrony than TD children. 2. Frequency Band Differences: In the θ frequency band, ASD children exhibited more long-range connections, while in the α frequency band, they showed fewer long-range connections, indicating that ASD children prefer local detail processing. 3. Brain Region Differences: Under positive emotional stimuli, the functional connectivity of the frontal lobe in ASD children was enhanced, while under negative emotional stimuli, the temporal lobe showed increased activity. 4. Classification Accuracy: The SVM classifier based on brain functional connectivity features achieved classification accuracies of 85% and 87% under positive and negative emotional stimuli, respectively, indicating that functional connectivity can serve as a potential biomarker for ASD diagnosis and classification.

Highlights of the Study

1. Analysis of Brain Functional Connectivity under Emotional Stimulation: This study is the first to systematically analyze the brain functional connectivity patterns of children with ASD under emotional stimuli, filling a gap in this field.
2. Multi-Frequency Band and Multi-Metric Analysis: The study used five frequency bands and four functional connectivity metrics to comprehensively assess the differences in brain functional connectivity between ASD and TD children.
3. High Classification Accuracy: By combining MRMR feature selection and SVM classifiers, the study achieved high classification accuracy, providing strong support for early ASD screening.
4. Application of Portable EEG Devices: The study used the portable EEG device Emotiv Epoc, demonstrating that even without advanced laboratory equipment, functional brain connectivity analysis has clinical application potential.

Significance and Value of the Study

This study provides new insights into understanding the neural mechanisms of ASD, especially the brain functional connectivity patterns during emotional processing. The results indicate that functional brain connectivity analysis can be an effective tool for ASD diagnosis and classification, particularly in early screening and personalized interventions. Additionally, the study demonstrates the potential of portable EEG devices in clinical research, laying the foundation for future research on larger samples and more diverse emotional stimuli.

Future Research Directions

Despite the significant progress made in this study, there are still some limitations. First, the study only used two types of emotional stimuli: positive and negative. Future research could expand to include more types of emotional stimuli, such as fear and anger. Second, the sample size was relatively small, and future studies could increase the sample size to improve the generalizability of the results. Finally, the study only used four functional connectivity metrics, and future research could introduce more types of connectivity metrics, such as time-frequency analysis and brain network modeling, to more comprehensively reveal the brain functional connectivity characteristics of ASD.