{"title":"Clinical validation of an abridged AIMS: Streamlining motor screening in the first-year infant","authors":"Teresa Fair-Field , Bharath Modayur","doi":"10.1016/j.earlhumdev.2025.106207","DOIUrl":null,"url":null,"abstract":"<div><div>This study validates an abridged version of the Alberta Infant Motor Scale (AIMS), termed the “salient set,” to streamline infant screening using video analysis and machine learning. Twenty-one retrospective infant videos were manually tagged by trained occupational therapists using only the 15-item salient set with support vector regressors (SVRs) trained on a larger sample (<em>n</em> = 102) predicting the true (full) AIMS score. The SVR demonstrated strong concurrent validity of the salient set with the full 58-item AIMS (Pearson correlation: 0.99). The abridged set showed high screening sensitivity (1.0) and specificity (0.895), while reducing evaluation time by 67 %. The salient set offers a useful contribution to machine learning by detecting an abridged set of items while still accurately and appropriately identifying infants for EI referral.</div></div>","PeriodicalId":11435,"journal":{"name":"Early human development","volume":"202 ","pages":"Article 106207"},"PeriodicalIF":2.2000,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Early human development","FirstCategoryId":"3","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0378378225000179","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"OBSTETRICS & GYNECOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
This study validates an abridged version of the Alberta Infant Motor Scale (AIMS), termed the “salient set,” to streamline infant screening using video analysis and machine learning. Twenty-one retrospective infant videos were manually tagged by trained occupational therapists using only the 15-item salient set with support vector regressors (SVRs) trained on a larger sample (n = 102) predicting the true (full) AIMS score. The SVR demonstrated strong concurrent validity of the salient set with the full 58-item AIMS (Pearson correlation: 0.99). The abridged set showed high screening sensitivity (1.0) and specificity (0.895), while reducing evaluation time by 67 %. The salient set offers a useful contribution to machine learning by detecting an abridged set of items while still accurately and appropriately identifying infants for EI referral.
期刊介绍：
Established as an authoritative, highly cited voice on early human development, Early Human Development provides a unique opportunity for researchers and clinicians to bridge the communication gap between disciplines. Creating a forum for the productive exchange of ideas concerning early human growth and development, the journal publishes original research and clinical papers with particular emphasis on the continuum between fetal life and the perinatal period; aspects of postnatal growth influenced by early events; and the safeguarding of the quality of human survival.
The first comprehensive and interdisciplinary journal in this area of growing importance, Early Human Development offers pertinent contributions to the following subject areas:
Fetology; perinatology; pediatrics; growth and development; obstetrics; reproduction and fertility; epidemiology; behavioural sciences; nutrition and metabolism; teratology; neurology; brain biology; developmental psychology and screening.
京公网安备 11010802042870号
京ICP备2023020795号-1
分享
求助内容：
应助结果提醒方式：
扫码关注我们
