不同意图场景眼动注视模式机器学习算法识别孤独症谱系障碍的研究

程蓉; 赵众; 侯文文; 周刚; 廖昊天; 张雪; 李晶

doi:10.7499/j.issn.1008-8830.2309073

PDF(2106 KB)

中国当代儿科杂志 ›› 2024, Vol. 26 ›› Issue (2) : 151-157. DOI: 10.7499/j.issn.1008-8830.2309073

论著·临床研究

不同意图场景眼动注视模式机器学习算法识别孤独症谱系障碍的研究

程蓉^1,2,3, 赵众³, 侯文文^1,2, 周刚³, 廖昊天³, 张雪³, 李晶^1,2

作者信息 +

Machine learning algorithms for identifying autism spectrum disorder through eye-tracking in different intention videos

CHENG Rong, ZHAO Zhong, HOU Wen-Wen, ZHOU Gang, LIAO Hao-Tian, ZHANG Xue, LI Jing

Author information +

文章历史 +

摘要

目的旨在探究孤独症谱系障碍（autism spectrum disorder, ASD）和正常发育（typically developing, TD）儿童在观看不同意图场景视频时的视觉感知差异，并通过机器学习算法，探索客观区分ASD和TD儿童的可行性。方法共纳入58名ASD儿童和50名TD儿童，观看包含联合意图和非联合意图的视频，使用儿童观看视频时在不同注视兴趣区的注视时长占比和注视次数占比作为原始特征指标，输入到分类器中构建分类模型，分析比较不同分类模型的分类准确率、灵敏度和特异度等指标。结果使用支持向量机、线性判别算法、决策树、随机森林和邻近算法（K=1、3、5、7）8种常见的分类器基于原始特征指标进行分类时，最高分类精度为81.90%。为进一步提高分类精度，采用特征重构方式，运用决策树分类器进行分类，分类模型的准确度高达91.43%，特异度为89.80%，灵敏度为92.86%，受试者操作特征曲线的曲线下面积为0.909（P<0.001）。结论采用眼动追踪数据构建的机器学习模型可精确地将ASD儿童与TD儿童区分开来，为开发快速客观筛查ASD的辅助工具提供科学依据。

Abstract

Objective To investigate the differences in visual perception between children with autism spectrum disorder (ASD) and typically developing (TD) children when watching different intention videos, and to explore the feasibility of machine learning algorithms in objectively distinguishing between ASD children and TD children. Methods A total of 58 children with ASD and 50 TD children were enrolled and were asked to watch the videos containing joint intention and non-joint intention, and the gaze duration and frequency in different areas of interest were used as original indicators to construct classifier-based models. The models were evaluated in terms of the indicators such as accuracy, sensitivity, and specificity. Results When using eight common classifiers, including support vector machine, linear discriminant analysis, decision tree, random forest, and K-nearest neighbors (with K values of 1, 3, 5, and 7), based on the original feature indicators, the highest classification accuracy achieved was 81.90%. A feature reconstruction approach with a decision tree classifier was used to further improve the accuracy of classification, and then the model showed the accuracy of 91.43%, the specificity of 89.80%, and the sensitivity of 92.86%, with an area under the receiver operating characteristic curve of 0.909 (P<0.001). Conclusions The machine learning model based on eye-tracking data can accurately distinguish ASD children from TD children, which provides a scientific basis for developing rapid and objective ASD screening tools.

导出引用

程蓉, 赵众, 侯文文, 周刚, 廖昊天, 张雪, 李晶. 不同意图场景眼动注视模式机器学习算法识别孤独症谱系障碍的研究[J]. 中国当代儿科杂志. 2024, 26(2): 151-157 https://doi.org/10.7499/j.issn.1008-8830.2309073

CHENG Rong, ZHAO Zhong, HOU Wen-Wen, ZHOU Gang, LIAO Hao-Tian, ZHANG Xue, LI Jing. Machine learning algorithms for identifying autism spectrum disorder through eye-tracking in different intention videos[J]. Chinese Journal of Contemporary Pediatrics. 2024, 26(2): 151-157 https://doi.org/10.7499/j.issn.1008-8830.2309073

参考文献

1 Zwaigenbaum L, Penner M. Autism spectrum disorder: advances in diagnosis and evaluation[J]. BMJ, 2018, 361: k1674. PMID: 29784657. DOI: 10.1136/bmj.k1674.
2 Falkmer T, Anderson K, Falkmer M, et al. Diagnostic procedures in autism spectrum disorders: a systematic literature review[J]. Eur Child Adolesc Psychiatry, 2013, 22(6): 329-340. PMID: 23322184. DOI: 10.1007/s00787-013-0375-0.
3 Li N, Chen G, Song X, et al. Prevalence of autism-caused disability among Chinese children: a national population-based survey[J]. Epilepsy Behav, 2011, 22(4): 786-789. PMID: 22079437. DOI: 10.1016/j.yebeh.2011.10.002.
4 Masi A, DeMayo MM, Glozier N, et al. An overview of autism spectrum disorder, heterogeneity and treatment options[J]. Neurosci Bull, 2017, 33(2): 183-193. PMID: 28213805. PMCID: PMC5360849. DOI: 10.1007/s12264-017-0100-y.
5 Hyman SL, Levy SE, Myers SM, et al. Identification, evaluation, and management of children with autism spectrum disorder[J]. Pediatrics, 2020, 145(1): e20193447. PMID: 31843864. DOI:10.1542/peds.2019-3447.
6 Shaw KA, Bilder DA, McArthur D, et al. Early identification of autism spectrum disorder among children aged 4 years: Autism and Developmental Disabilities Monitoring Network, 11 Sites, United States, 2020[J]. MMWR Surveill Summ, 2023, 72(1): 1-15. PMID: 36952289. PMCID: PMC10042615. DOI: 10.15585/mmwr.ss7201a1.
7 Mandell DS, Novak MM, Zubritsky CD. Factors associated with age of diagnosis among children with autism spectrum disorders[J]. Pediatrics, 2005, 116(6): 1480-1486. PMID: 16322174. PMCID: PMC2861294. DOI: 10.1542/peds.2005-0185.
8 Dalton JC, Crais ER, Velleman SL. Joint attention and oromotor abilities in young children with and without autism spectrum disorder[J]. J Commun Disord, 2017, 69: 27-43. PMID: 28704690. DOI: 10.1016/j.jcomdis.2017.06.002.
9 Woo BM, Spelke ES. Infants and toddlers leverage their understanding of action goals to evaluate agents who help others[J]. Child Dev, 2023, 94(3): 734-751. PMID: 36752158. DOI: 10.1111/cdev.13895.
10 Tomasello M, Carpenter M, Call J, et al. Understanding and sharing intentions: the origins of cultural cognition[J]. Behav Brain Sci, 2005, 28(5): 675-691. PMID: 16262930. DOI: 10.1017/S0140525X05000129.
11 Fawcett C, Kreutz G. Twelve-month-old infants' physiological responses to music are affected by others' positive and negative reactions[J]. Infancy, 2021, 26(6): 784-797. PMID: 34120402. DOI: 10.1111/infa.12415.
12 Ulber J, Hamann K, Tomasello M. How 18- and 24-month-old peers divide resources among themselves[J]. J Exp Child Psychol, 2015, 140: 228-244. PMID: 26283235. DOI: 10.1016/j.jecp.2015.07.009.
13 Fawcett C, Gredeb?ck G. Infants use social context to bind actions into a collaborative sequence[J]. Dev Sci, 2013, 16(6): 841-849. PMID: 24118711. DOI: 10.1111/desc.12074.
14 Fawcett C, Tun?gen? B. Infants' use of movement synchrony to infer social affiliation in others[J]. J Exp Child Psychol, 2017, 160: 127-136. PMID: 28427721. DOI: 10.1016/j.jecp.2017.03.014.
15 Fawcett C, Liszkowski U. Observation and initiation of joint action in infants[J]. Child Dev, 2012, 83(2): 434-441. PMID: 22277061. DOI: 10.1111/j.1467-8624.2011.01717.x.
16 Fitzpatrick P, Romero V, Amaral JL, et al. Social motor synchronization: insights for understanding social behavior in autism[J]. J Autism Dev Disord, 2017, 47(7): 2092-2107. PMID: 28425022. DOI: 10.1007/s10803-017-3124-2.
17 Kaliukhovich DA, Manyakov NV, Bangerter A, et al. Social attention to activities in children and adults with autism spectrum disorder: effects of context and age[J]. Mol Autism, 2020, 11(1): 79. PMID: 33076994. PMCID: PMC7574440. DOI: 10.1186/s13229-020-00388-5.
18 Frazier TW, Strauss M, Klingemier EW, et al. A meta-analysis of gaze differences to social and nonsocial information between individuals with and without autism[J]. J Am Acad Child Adolesc Psychiatry, 2017, 56(7): 546-555. PMID: 28647006. PMCID: PMC5578719. DOI: 10.1016/j.jaac.2017.05.005.
19 侯文文, 李晶, 李婷玉, 等. 孤独症谱系障碍患者社会性注意异常的机制[J]. 中国科学（生命科学）, 2019, 49(1): 59-66. DOI: 10.1360/N052018-00151.
20 李晶, 朱莉琪. 高功能孤独症儿童的合作行为[J]. 心理学报, 2014, 46(9): 1301-1316. DOI: 10.3724/SP.J.1041.2014.01301.
21 Liebal K, Colombi C, Rogers SJ, et al. Helping and cooperation in children with autism[J]. J Autism Dev Disord, 2008, 38(2): 224-238. PMID: 17694374. PMCID: PMC2758368. DOI: 10.1007/s10803-007-0381-5.
22 Helminen TM, Lepp?nen JM, Eriksson K, et al. Atypical physiological orienting to direct gaze in low-functioning children with autism spectrum disorder[J]. Autism Res, 2017, 10(5): 810-820. PMID: 28244277. DOI: 10.1002/aur.1738.
23 Alca?iz M, Chicchi-Giglioli IA, Carrasco-Ribelles LA, et al. Eye gaze as a biomarker in the recognition of autism spectrum disorder using virtual reality and machine learning: a proof of concept for diagnosis[J]. Autism Res, 2022, 15(1): 131-145. PMID: 34811930. DOI: 10.1002/aur.2636.
24 Kong XJ, Wei Z, Sun B, et al. Different eye tracking patterns in autism spectrum disorder in toddler and preschool children[J]. Front Psychiatry, 2022, 13: 899521. PMID: 35757211. PMCID: PMC9218189. DOI: 10.3389/fpsyt.2022.899521.
25 Wan G, Kong X, Sun B, et al. Applying eye tracking to identify autism spectrum disorder in children[J]. J Autism Dev Disord, 2019, 49(1): 209-215. PMID: 30097760. DOI: 10.1007/s10803-018-3690-y.
26 Jiang M, Francis SM, Srishyla D, et al. Classifying individuals with ASD through facial emotion recognition and eye-tracking[J]. Annu Int Conf IEEE Eng Med Biol Soc, 2019, 2019: 6063-6068. PMID: 31947228. DOI: 10.1109/EMBC.2019.8857005.
27 Kanhirakadavath MR, Chandran MSM. Investigation of eye-tracking scan path as a biomarker for autism screening using machine learning algorithms[J]. Diagnostics (Basel), 2022, 12(2): 518. PMID: 35204608. PMCID: PMC8871384. DOI: 10.3390/diagnostics12020518.
28 Liu W, Li M, Yi L. Identifying children with autism spectrum disorder based on their face processing abnormality: a machine learning framework[J]. Autism Res, 2016, 9(8): 888-898. PMID: 27037971. DOI: 10.1002/aur.1615.
29 Minissi ME, Chicchi Giglioli IA, Mantovani F, et al. Assessment of the autism spectrum disorder based on machine learning and social visual attention: a systematic review[J]. J Autism Dev Disord, 2022, 52(5): 2187-2202. PMID: 34101081. PMCID: PMC9021060. DOI: 10.1007/s10803-021-05106-5.
30 Wei Q, Cao H, Shi Y, et al. Machine learning based on eye-tracking data to identify autism spectrum disorder: a systematic review and meta-analysis[J]. J Biomed Inform, 2023, 137: 104254. PMID: 36509416. DOI: 10.1016/j.jbi.2022.104254.
31 Zhao Z, Tang H, Zhang X, et al. Classification of children with autism and typical development using eye-tracking data from face-to-face conversations: machine learning model development and performance evaluation[J]. J Med Internet Res, 2021, 23(8): e29328. PMID: 34435957. PMCID: PMC8440949. DOI: 10.2196/29328.
32 Zhao Z, Wei J, Xing J, et al. Use of oculomotor behavior to classify children with autism and typical development: a novel implementation of the machine learning approach[J]. J Autism Dev Disord, 2023, 53(3): 934-946. PMID: 35913654. DOI: 10.1007/s10803-022-05685-x.
33 李西, 田真玲, 姜孟. 机器学习在孤独症筛查与诊断中的应用研究进展[J]. 中国特殊教育, 2022(9): 47-54. DOI: 10.3969/j.issn.1007-3728.2022.09.006.
34 American Psychiatric Association, DSM-5 Task Force. Diagnostic and Statistical Manual of Mental Disorders: DSM-5[M]. 5th ed. Washington: American Psychiatric Association, 2013: 36-37.
35 张厚粲, 王晓平. 瑞文标准推理测验在我国的修订[J]. 心理学报, 1989, 21(2): 113-121.
36 Kang J, Han X, Song J, et al. The identification of children with autism spectrum disorder by SVM approach on EEG and eye-tracking data[J]. Comput Biol Med, 2020, 120: 103722. PMID: 32250854. DOI: 10.1016/j.compbiomed.2020.103722.
37 康健楠, 韩晓雅, 耿新玲, 等. 基于眼动追踪技术的孤独症谱系障碍儿童面孔异常加工识别研究[J]. 中国康复医学杂志, 2022, 37(7): 933-936. DOI: 10.3969/j.issn.1001-1242.2022.07.012.
38 杨剑锋, 乔佩蕊, 李永梅, 等. 机器学习分类问题及算法研究综述[J]. 统计与决策, 2019, 35(6): 36-40. DOI: 10.13546/j.cnki.tjyjc.2019.06.008.
39 张润, 王永滨. 机器学习及其算法和发展研究[J]. 中国传媒大学学报（自然科学版）, 2016, 23(2): 10-18. DOI: 10.3969/j.issn.1673-4793.2016.02.002.
40 Chita-Tegmark M. Social attention in ASD: a review and meta-analysis of eye-tracking studies[J]. Res Dev Disabil, 2016, 48: 79-93. PMID: 26547134. DOI: 10.1016/j.ridd.2015.10.011.
41 Yaneva V, Ha LA, Eraslan S, et al. Detecting high-functioning autism in adults using eye tracking and machine learning[J]. IEEE Trans Neural Syst Rehabil Eng, 2020, 28(6): 1254-1261. PMID: 32356755. DOI: 10.1109/TNSRE.2020.2991675.