Machine Learning Identifies Infant Speech Preference as an Important Predictor for Early Autism Risk Detection

Jiang, Yuhan; Fu, Qiang

doi:https://doi.org/10.18910/104119

Machine Learning Identifies Infant Speech Preference as an Important Predictor for Early Autism Risk Detection

Jiang, Yuhan; Fu, Qiang

The University of Osaka - Tongji University Joint International and Interdisciplinary Forum: Toward the Integration of Society and Knowledge Official Forum Proceedings, 2026, 41-56

アクセス数:120件（2026-10-19 01:22 集計）

固定URL: https://doi.org/10.18910/104123

閲覧可能ファイル

ファイル	フォーマット	利用条件	サイズ	閲覧回数	利用開始日	説明	information
tisk_proceedings_041	pdf	なし	497 KB	77	2026.03.25

論文情報

ファイル出力 EndNote Basic出力 Mendeley出力

タイトル	Machine Learning Identifies Infant Speech Preference as an Important Predictor for Early Autism Risk Detection
著者	著者名 Jiang, Yuhan
著者	著者名 Fu, Qiang
キーワード等	Machine learning
	Early autistic risk identification
	Multimodal behaviors
	Speech preference
抄録	While conventional ASD risk assessments rely on later-emerging behavioral features, early vocal preferences offer infant-available predictive markers. Therefore, multimodal measures integrating behavioral features with vocal preference may enhance autistic risk prediction accuracy. This study developed machine learning models to identify high autistic risk infants using multimodal measures, including gaze duration of sound contrasts (speech vs. non-vocal; speech vs. monkey calls) and learning/motor scores. Analysis of 9/12 month infant data (56 high-risk, 82 low-risk) showed SVM performed best after optimization (71.43% accuracy, AUC=0.46). Average model feature importance analysis identified that key predictors are 9-month human-over-monkey preference, 9-month speech-over-nonspeech preference, and 12-month receptive language percentiles. Findings demonstrate that gaze-based speech preference assessment addresses critical limitations of traditional behavioral markers, and when integrated with predictive modeling, opens new pathways for significant advances in early autism risk detection.
出版者	Fellowship for Integration of Knowledge with Society, The University of Osaka
収録物名	The University of Osaka - Tongji University Joint International and Interdisciplinary Forum: Toward the Integration of Society and Knowledge Official Forum Proceedings
ページ	41-56
刊行年月	2026-03
言語	英語
ハンドルURL	https://hdl.handle.net/11094/104123
DOI	https://doi.org/10.18910/104123
関連情報 (isPartOf)	DOI (出版社版) https://doi.org/10.18910/104119
アクセス権	open access
出版タイプ	Version of Record
カテゴリ	会議発表論文 Conference Paper
カテゴリ	The University of Osaka - Tongji University Joint International and Interdisciplinary Forum: Toward the Integration of Society and Knowledge Official Forum Proceedings

資源タイプ	会議発表論文
ローカル資源タイプ	会議発表論文
DCMI資源タイプ	text
DCTERMS.bibliographicCitation	The University of Osaka - Tongji University Joint International and Interdisciplinary Forum: Toward the Integration of Society and Knowledge Official Forum Proceedings. 2026, p. 41-56
DC.title	Machine Learning Identifies Infant Speech Preference as an Important Predictor for Early Autism Risk Detection
DC.creator	Jiang, Yuhan
DC.creator	Fu, Qiang
DC.publisher	Fellowship for Integration of Knowledge with Society, The University of Osaka
DC.language' scheme='DCTERMS.RFC1766	英語
DC.type' scheme='DCTERMS.DCMIType	text
DCTERMS.issued' scheme='DCTERMS.W3CDTF	2026-03
DC.identifier	https://doi.org/10.18910/104119
DC.subject	Machine learning
	Early autistic risk identification
	Multimodal behaviors
	Speech preference
DCTERMS.abstract	While conventional ASD risk assessments rely on later-emerging behavioral features, early vocal preferences offer infant-available predictive markers. Therefore, multimodal measures integrating behavioral features with vocal preference may enhance autistic risk prediction accuracy. This study developed machine learning models to identify high autistic risk infants using multimodal measures, including gaze duration of sound contrasts (speech vs. non-vocal; speech vs. monkey calls) and learning/motor scores. Analysis of 9/12 month infant data (56 high-risk, 82 low-risk) showed SVM performed best after optimization (71.43% accuracy, AUC=0.46). Average model feature importance analysis identified that key predictors are 9-month human-over-monkey preference, 9-month speech-over-nonspeech preference, and 12-month receptive language percentiles. Findings demonstrate that gaze-based speech preference assessment addresses critical limitations of traditional behavioral markers, and when integrated with predictive modeling, opens new pathways for significant advances in early autism risk detection.
DC.identifier	10.18910/104123
DC.format	104123
citation_title	Machine Learning Identifies Infant Speech Preference as an Important Predictor for Early Autism Risk Detection
citation_author	Jiang, Yuhan
citation_author	Fu, Qiang
citation_publisher	Fellowship for Integration of Knowledge with Society, The University of Osaka
citation_language	英語
citation_date	2026-03
citation_journal_title	The University of Osaka - Tongji University Joint International and Interdisciplinary Forum: Toward the Integration of Society and Knowledge Official Forum Proceedings
citation_firstpage	41
citation_lastpage	56
citation_doi	https://doi.org/10.18910/104119
citation_public_url	https://hdl.handle.net/11094/104123
citation_doi_detail	10.18910/104123
citation_doi	https://doi.org/10.18910/104123

Machine Learning Identifies Infant Speech Preference as an Important Predictor for Early Autism Risk Detection

閲覧可能ファイル

論文情報

論文詳細を表示