一种用于幼儿自闭症筛查的机器学习策略。
A Machine Learning Strategy for Autism Screening in Toddlers.
机构信息
Departments of Chemical Engineering.
Psychology.
出版信息
J Dev Behav Pediatr. 2019 Jun;40(5):369-376. doi: 10.1097/DBP.0000000000000668.
OBJECTIVE
Autism spectrum disorder (ASD) screening can improve prognosis via early diagnosis and intervention, but lack of time and training can deter pediatric screening. The Modified Checklist for Autism in Toddlers, Revised (M-CHAT-R) is a widely used screener but requires follow-up questions and error-prone human scoring and interpretation. We consider an automated machine learning (ML) method for overcoming barriers to ASD screening, specifically using the feedforward neural network (fNN).
METHODS
The fNN technique was applied using archival M-CHAT-R data of 14,995 toddlers (age 16-30 months, 46.51% male). The 20 M-CHAT-R items were inputs, and ASD diagnosis after follow-up and diagnostic evaluation (i.e., ASD or not ASD) was the output. The sample was divided into subgroups by race (i.e., white and black), sex (i.e., boys and girls), and maternal education (i.e., below and above 15 years of education completed) to examine subgroup differences. Each subgroup was evaluated for best-performing fNN models.
RESULTS
For the total sample, best results yielded 99.72% correct classification using 18 items. Best results yielded 99.92% correct classification using 14 items for white toddlers and 99.79% correct classification using 18 items for black toddlers. In boys, best results yielded 99.64% correct classification using 18 items, whereas best results yielded 99.95% correct classification using 18 items in girls. For the case when maternal education is 15 years or less (i.e., associate degree and below), best results were 99.75% correct classification when using 16 items. Results were essentially the same when maternal education was 16 years or more (i.e., above associate degree); that is, 99.70% correct classification was obtained using 16 items.
CONCLUSION
The ML method was comparable to the M-CHAT-R with follow-up items in accuracy of ASD diagnosis while using fewer items. Therefore, ML may be a beneficial tool in implementing automatic, efficient scoring that negates the need for labor-intensive follow-up and circumvents human error, providing an advantage over previous screening methods.
目的
自闭症谱系障碍（ASD）筛查可以通过早期诊断和干预来改善预后，但缺乏时间和培训可能会阻碍儿科筛查。改良婴幼儿自闭症检查表（M-CHAT-R）是一种广泛使用的筛查工具，但需要后续问题和容易出错的人工评分和解释。我们考虑使用自动化机器学习（ML）方法来克服 ASD 筛查的障碍，特别是使用前馈神经网络（fNN）。
方法
使用存档的 14995 名 16-30 个月大的幼儿（46.51%为男性）的 M-CHAT-R 数据应用 fNN 技术。20 项 M-CHAT-R 项目为输入，随访和诊断评估后的 ASD 诊断（即 ASD 或非 ASD）为输出。根据种族（即白人和黑人）、性别（即男孩和女孩）和母亲教育程度（即完成教育程度低于或高于 15 年）将样本分为亚组，以检查亚组差异。对每个亚组进行最佳 fNN 模型评估。
结果
对于总样本，使用 18 项最佳结果可实现 99.72%的正确分类。对于白人幼儿，使用 14 项最佳结果可实现 99.92%的正确分类，对于黑人幼儿，使用 18 项最佳结果可实现 99.79%的正确分类。在男孩中，使用 18 项最佳结果可实现 99.64%的正确分类，而在女孩中，使用 18 项最佳结果可实现 99.95%的正确分类。对于母亲教育程度为 15 年或以下（即副学士学位及以下）的情况，使用 16 项最佳结果可实现 99.75%的正确分类。当母亲教育程度为 16 年或以上（即副学士学位以上）时，结果基本相同，即使用 16 项可获得 99.70%的正确分类。
结论
ML 方法在 ASD 诊断的准确性上与使用后续项目的 M-CHAT-R 相当，同时使用的项目更少。因此，ML 可能是一种有益的工具，可以实现自动、高效的评分，从而无需劳动密集型的后续工作，并避免人为错误，相对于以前的筛查方法具有优势。
Zotero 插件