Effective Autism Classification Through Grasping Kinematics

• Published in: Autism research Vol. 18; no. 6; pp. 1170 - 1181
• Main Authors: Freud, Erez, Ahmad, Zoha, Shelef, Eitan, Hadad, Bat Sheva
• Format: Journal Article
• Language: English
• Published: Hoboken, USA John Wiley & Sons, Inc 01.06.2025
• Subjects: Adult; autism; Autistic Disorder - classification; Autistic Disorder - diagnosis; Autistic Disorder - physiopathology; Biomechanical Phenomena - physiology; Female; grasping; Hand Strength - physiology; Humans; machine learning; Male; motion tracking; Psychomotor Performance - physiology; Reproducibility of Results; visuomotor control; Young Adult; grasping; visuomotor control; motion tracking; machine learning; autism
• ISSN: 1939-3792; 1939-3806
• DOI: 10.1002/aur.70049

ABSTRACT Autism is a complex neurodevelopmental condition, where motor abnormalities play a central role alongside social and communication difficulties. These motor symptoms often manifest in early childhood, making them critical targets for early diagnosis and intervention. This study aimed to assess whether kinematic features from a naturalistic grasping task could accurately distinguish autistic participants from non‐autistic ones. We analyzed grasping movements of autistic and non‐autistic young adults, tracking two markers placed on the thumb and index finger. Using a subject‐wise cross‐validated classifiers, we achieved accuracy scores of above 84%. Receiver operating characteristic analysis revealed strong classification performance with area under the curve values of above 0.95 at the subject‐wise analysis and above 0.85 at the trial‐wise analysis. These findings indicate strong reliability in accurately distinguishing autistic participants from non‐autistic ones. These findings suggest that subtle motor control differences can be effectively captured, offering a promising approach for developing accessible and reliable diagnostic tools for autism.