Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Oct 11:17:1236637.
doi: 10.3389/fnins.2023.1236637. eCollection 2023.

Identification of diagnostic markers for ASD: a restrictive interest analysis based on EEG combined with eye tracking

Binbin Sun  1 , Bryan Wang  2   3 , Zhen Wei  1 , Zhe Feng  1 , Zhi-Liu Wu  1 , Walid Yassin  2   4   5 , William S Stone  5 , Yan Lin  1 , Xue-Jun Kong  2   5
Affiliations

Identification of diagnostic markers for ASD: a restrictive interest analysis based on EEG combined with eye tracking

Binbin Sun et al. Front Neurosci. .

Abstract

Electroencephalography (EEG) functional connectivity (EFC) and eye tracking (ET) have been explored as objective screening methods for autism spectrum disorder (ASD), but no study has yet evaluated restricted and repetitive behavior (RRBs) simultaneously to infer early ASD diagnosis. Typically developing (TD) children (n = 27) and ASD (n = 32), age- and sex-matched, were evaluated with EFC and ET simultaneously, using the restricted interest stimulus paradigm. Network-based machine learning prediction (NBS-predict) was used to identify ASD. Correlations between EFC, ET, and Autism Diagnostic Observation Schedule-Second Edition (ADOS-2) were performed. The Area Under the Curve (AUC) of receiver-operating characteristics (ROC) was measured to evaluate the predictive performance. Under high restrictive interest stimuli (HRIS), ASD children have significantly higher α band connectivity and significantly more total fixation time (TFT)/pupil enlargement of ET relative to TD children (p = 0.04299). These biomarkers were not only significantly positively correlated with each other (R = 0.716, p = 8.26e-4), but also with ADOS total scores (R = 0.749, p = 34e-4) and RRBs sub-score (R = 0.770, p = 1.87e-4) for EFC (R = 0.641, p = 0.0148) for TFT. The accuracy of NBS-predict in identifying ASD was 63.4%. ROC curve demonstrated TFT with 91 and 90% sensitivity, and 78.7% and 77.4% specificity for ADOS total and RRB sub-scores, respectively. Simultaneous EFC and ET evaluation in ASD is highly correlated with RRB symptoms measured by ADOS-2. NBS-predict of EFC offered a direct prediction of ASD. The use of both EFC and ET improve early ASD diagnosis.

Keywords: ASD biomarker; ASD early diagnosis; EEG; eye tracking; functional connectivity.

PubMed Disclaimer

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Figures

Figure 1
Figure 1
Restricted interest objects.
Figure 2
Figure 2
Stimulus flowchart.
Figure 3
Figure 3
10–20 electrode distribution map of the EEG system. (A) Original channel locations distribution map; (B) Channel locations distribution map after conversion.
Figure 4
Figure 4
The α-band WPLI of ASD and TD under the LRIS (low) and HRIS (high). (A) ASD group alpha band WPLI r matrix graph and topology graph; (B) TD group alpha band WPLI r matrix graph and topology graph; (C) Brain regions with statistical difference in WPLI in ASD group; (D) Brain regions with statistical difference in WPLI in TD group. ASD, Autism Spectrum Disorder; TD, Typically developing children; High, high restrictive interest stimuli (HRIS); Low, low restrictive interest stimuli (LRIS). The first behavior matrix graph and the second behavior topology graph both represent the same information, and the surrounding numbers represent the energy range, which is the result graph of the functional connectivity analysis.
Figure 5
Figure 5
NBS-predict result. (A) Connected brain map; (B) Connected circle map; (C) Adjacency matrix; (D) Classification performance; 1 represents the ASD group, 0 represents the TD group. C, central region; P, parietal region; T, temporal region; F, frontal region; z, midline position of the brain.
Figure 6
Figure 6
Correlation between alpha-band WPLI value and pupil size in ASD and TD under the LRIS (low) and HRIS (high). (A) Correlation between alpha-band WPLI value and pupil size in ASD under the HRIS; (B) Correlation between alpha-band WPLI value and pupil size in TD under the HRIS; (C) Correlation between alpha-band WPLI value and pupil size in ASD under the LRIS; (D) Correlation between alpha-band WPLI value and pupil size in TD under the LRIS. ASD, Autism Spectrum Disorder; TD, Typically developing children; low, low restricted interest stimuli; high, high restricted interest stimuli.
Figure 7
Figure 7
Correlation between WPLI value of α frequency band and TFT in ASD group and TD group under the conditions of LRIS (low) and HRIS (high). (A) Correlation between alpha-band WPLI value and TFT in ASD under the HRIS; (B) Correlation between alpha-band WPLI value and TFT in TD under the HRIS; (C) Correlation between alpha-band WPLI value and TFT in ASD under the LRIS; (D) Correlation between alpha-band WPLI value and TFT in TD under the LRIS. TFT, total fixation time; ASD, Autism Spectrum Disorder; TD, Typically developing children; low, low restricted interest stimuli (LRIS); high, high restricted interest stimuli (HRIS).
Figure 8
Figure 8
Correlation between TFT and ADOS total scores in children with ASD under HRIS (high). TFT, total fixation time; HRIS, high restricted interest stimuli; ADOS score is ADOS-2 total score.
Figure 9
Figure 9
Correlation between TFT and RRB sub-scores in children with ASD under HRIS (high). TFT, total fixation time; RRB, restricted repetitive behavior; HRIS, high restricted interest stimuli; Restricted behavior is RRB sub-scores of ADOS-2.
Figure 10
Figure 10
Correlations between WPLI value of α frequency band, ADOS score and restricted behavior in ASD group under the conditions of LRIS (low) and HRIS (high). (A) Correlations between WPLI value of α frequency band and ADOS score under the conditions of HRIS; (B) Correlations between WPLI value of α frequency band and ADOS score under the conditions of LRIS; (C) Correlations between WPLI value of α frequency band and RRBS score under the conditions of HRIS; (D) Correlations between WPLI value of α frequency band and RRBS score under the conditions of LRIS. Low, low restricted interest stimuli (LRIS); high, high restricted interest stimuli (HRIS). ADOS score is ADOS-2 total score, Restricted behavior is RRB sub-scores of ADOS-2.
Figure 11
Figure 11
ROC curves of TFT, ADOS total and ADOS-RRB in ASD children under HRIS (high). TFT, Total fixation time; ADOS total, ADOS-2 total scores; ADOS-RRB, Restricted repetitive behavior sub-scores of ADOS-2; HRIS, high restricted interest stimuli.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Ahmadlou M., Adeli H., Adeli A. (2012). Fuzzy synchronization likelihood-wavelet methodology for diagnosis of autism spectrum disorder. J. Neurosci. Methods 211, 203–209. doi: 10.1016/j.jneumeth.2012.08.020 - DOI - PubMed
    1. Alotaibi N., Maharatna K. (2021). Classification of autism Spectrum disorder from EEG-based functional brain connectivity analysis. Neural Comput. 33, 1914–1941. doi: 10.1162/neco_a_01394 - DOI - PubMed
    1. Anthony L. G., Kenworthy L., Yerys B. E., Jankowski K. F., James J. D., Harms M. B., et al. . (2013). Lnlcresls in high-unclioning aulism arc morc intensc, intcrlcring and idiosyncratic than thosc in ncurolypical development. Dev. Psychopathol. 25:643 652. doi: 10.1017/S0954579413000072 - DOI - PMC - PubMed
    1. Artoni P., Piffer A., Vinci V., LeBlanc J., Nelson C. A., Hensch T. K., et al. . (2020). Deep learning of spontaneous arousal fluctuations detects early cholinergic defects across neurodevelopmental mouse models and patients. Proc. Natl. Acad. Sci. U. S. A. 117, 23298–23303. doi: 10.1073/pnas.1820847116 - DOI - PMC - PubMed
    1. Baccino T., Manunta Y. (2006). Eye-fixation related potentials: insight into parafoveal processing. Int. J. Psychophysiol. 19, 204–215.

Grants and funding

This research was funded by General Projects of Shenzhen Science and Technology Innovation Committee, China #JCYJ20210324131211031, Key-Area Research and development program of Guangdong Province 2019B030335001, Shenzhen Natural Science Foundation (JCYJ20190806142412826), and Massachusetts General Hospital #233263, and Shenzhen Sanming project (SZSM201512009) and Shenzhen Maternity and Child Healthcare Hospital (FYA2022018).