DASD- diagnosing autism spectrum disorder based on stereotypical hand-flapping movements using multi-stream neural networks and attention mechanisms

IntroductionThe early detection and diagnosis of autism spectrum disorder (ASD) remain critical challenges in developmental healthcare, with traditional diagnostic methods relying heavily on subjective clinical observations.MethodsIn this paper, we introduce an innovative multi-stream framework that...

Full description

Saved in:
Bibliographic Details
Main Authors: Theyazn H. H. Aldhyani, Abdullah H. Al-Nefaie
Format: Article
Language:English
Published: Frontiers Media S.A. 2025-07-01
Series:Frontiers in Physiology
Subjects:
Online Access:https://www.frontiersin.org/articles/10.3389/fphys.2025.1593965/full
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850086140797779968
author Theyazn H. H. Aldhyani
Theyazn H. H. Aldhyani
Abdullah H. Al-Nefaie
Abdullah H. Al-Nefaie
author_facet Theyazn H. H. Aldhyani
Theyazn H. H. Aldhyani
Abdullah H. Al-Nefaie
Abdullah H. Al-Nefaie
author_sort Theyazn H. H. Aldhyani
collection DOAJ
description IntroductionThe early detection and diagnosis of autism spectrum disorder (ASD) remain critical challenges in developmental healthcare, with traditional diagnostic methods relying heavily on subjective clinical observations.MethodsIn this paper, we introduce an innovative multi-stream framework that seamlessly integrates three state-of-the-art convolutional neural networks, namely, EfficientNetV2B0, ResNet50V2, DenseNet121, and Multi-Stream models to analyze stereotypical movements, particularly hand-flapping behaviors automatically. Our architecture incorporates sophisticated spatial and temporal attention mechanisms enhanced by hierarchical feature fusion and adaptive temporal sampling techniques designed to extract characteristics of ASD related movements across multiple scales. The system includes a custom designed temporal attention module that effectively captures the rhythmic nature of hand-flapping behaviors. The spatial attention mechanisms method was used to enhance the proposed models by focusing on the movement characteristics of the patients in the video. The experimental validation was conducted using the Self-Stimulatory Behavior Dataset (SSBD), which includes 66 videos.ResultsThe Multi-Stream framework demonstrated exceptional performance, with 96.55% overall accuracy, 100% specificity, and 94.12% sensitivity in terms of hand-flapping detection and an impressive F1 score of 97%.DiscussionThis research can provide healthcare professionals with a reliable, automated tool for early ASD screening that offers objective, quantifiable metrics that complement traditional diagnostic methods.
format Article
id doaj-art-bd0634b048524194b472c80d23aae898
institution DOAJ
issn 1664-042X
language English
publishDate 2025-07-01
publisher Frontiers Media S.A.
record_format Article
series Frontiers in Physiology
spelling doaj-art-bd0634b048524194b472c80d23aae8982025-08-20T02:43:33ZengFrontiers Media S.A.Frontiers in Physiology1664-042X2025-07-011610.3389/fphys.2025.15939651593965DASD- diagnosing autism spectrum disorder based on stereotypical hand-flapping movements using multi-stream neural networks and attention mechanismsTheyazn H. H. Aldhyani0Theyazn H. H. Aldhyani1Abdullah H. Al-Nefaie2Abdullah H. Al-Nefaie3King Salman Center for Disability Research, Riyadh, Saudi ArabiaApplied college in Abqaiq, King Faisal University, Al-Ahsa, Saudi ArabiaKing Salman Center for Disability Research, Riyadh, Saudi ArabiaDepartment of Quantitative Methods, School of Business, King Faisal University, Al-Ahsa, Saudi ArabiaIntroductionThe early detection and diagnosis of autism spectrum disorder (ASD) remain critical challenges in developmental healthcare, with traditional diagnostic methods relying heavily on subjective clinical observations.MethodsIn this paper, we introduce an innovative multi-stream framework that seamlessly integrates three state-of-the-art convolutional neural networks, namely, EfficientNetV2B0, ResNet50V2, DenseNet121, and Multi-Stream models to analyze stereotypical movements, particularly hand-flapping behaviors automatically. Our architecture incorporates sophisticated spatial and temporal attention mechanisms enhanced by hierarchical feature fusion and adaptive temporal sampling techniques designed to extract characteristics of ASD related movements across multiple scales. The system includes a custom designed temporal attention module that effectively captures the rhythmic nature of hand-flapping behaviors. The spatial attention mechanisms method was used to enhance the proposed models by focusing on the movement characteristics of the patients in the video. The experimental validation was conducted using the Self-Stimulatory Behavior Dataset (SSBD), which includes 66 videos.ResultsThe Multi-Stream framework demonstrated exceptional performance, with 96.55% overall accuracy, 100% specificity, and 94.12% sensitivity in terms of hand-flapping detection and an impressive F1 score of 97%.DiscussionThis research can provide healthcare professionals with a reliable, automated tool for early ASD screening that offers objective, quantifiable metrics that complement traditional diagnostic methods.https://www.frontiersin.org/articles/10.3389/fphys.2025.1593965/fullautism spectrum disorderdeep learningstereotypical movementshandflapping detectionmulti-stream architectureattention mechanisms
spellingShingle Theyazn H. H. Aldhyani
Theyazn H. H. Aldhyani
Abdullah H. Al-Nefaie
Abdullah H. Al-Nefaie
DASD- diagnosing autism spectrum disorder based on stereotypical hand-flapping movements using multi-stream neural networks and attention mechanisms
Frontiers in Physiology
autism spectrum disorder
deep learning
stereotypical movements
handflapping detection
multi-stream architecture
attention mechanisms
title DASD- diagnosing autism spectrum disorder based on stereotypical hand-flapping movements using multi-stream neural networks and attention mechanisms
title_full DASD- diagnosing autism spectrum disorder based on stereotypical hand-flapping movements using multi-stream neural networks and attention mechanisms
title_fullStr DASD- diagnosing autism spectrum disorder based on stereotypical hand-flapping movements using multi-stream neural networks and attention mechanisms
title_full_unstemmed DASD- diagnosing autism spectrum disorder based on stereotypical hand-flapping movements using multi-stream neural networks and attention mechanisms
title_short DASD- diagnosing autism spectrum disorder based on stereotypical hand-flapping movements using multi-stream neural networks and attention mechanisms
title_sort dasd diagnosing autism spectrum disorder based on stereotypical hand flapping movements using multi stream neural networks and attention mechanisms
topic autism spectrum disorder
deep learning
stereotypical movements
handflapping detection
multi-stream architecture
attention mechanisms
url https://www.frontiersin.org/articles/10.3389/fphys.2025.1593965/full
work_keys_str_mv AT theyaznhhaldhyani dasddiagnosingautismspectrumdisorderbasedonstereotypicalhandflappingmovementsusingmultistreamneuralnetworksandattentionmechanisms
AT theyaznhhaldhyani dasddiagnosingautismspectrumdisorderbasedonstereotypicalhandflappingmovementsusingmultistreamneuralnetworksandattentionmechanisms
AT abdullahhalnefaie dasddiagnosingautismspectrumdisorderbasedonstereotypicalhandflappingmovementsusingmultistreamneuralnetworksandattentionmechanisms
AT abdullahhalnefaie dasddiagnosingautismspectrumdisorderbasedonstereotypicalhandflappingmovementsusingmultistreamneuralnetworksandattentionmechanisms