5 February 2024 in IEEE Transactions on Neural Systems and Rehabilitation Engineering
Autism spectrum disorder (ASD) one of the fastest-growing diseases in the world is a group of neurodevelopmental disorders. Eye movement as a biomarker and clinical manifestation represents unconscious brain processes that can objectively disclose abnormal eye fixation of ASD. With the aid of eye-tracking technology, plentiful methods that identify ASD based on eye movements have been developed, but there are rarely works specifically for scanpaths. Scanpaths as visual representations describe eye movement dynamics on stimuli. In this paper, we propose a scanpath-based ASD detection method, which aims to learn the atypical visual pattern of ASD through continuous dynamic changes in gaze distribution. We extract four sequence features from scanpaths that represent changes and the differences in feature space and gaze behavior patterns between ASD and typical development (TD) are explored based on two similarity measures, multimatch and dynamic time warping (DTW). It indicates that ASD children show more individual specificity, while normal children tend to develop similar visual patterns. The most noticeable contrasts lie in the duration of attention and the spatial distribution of visual attention along the vertical direction. Classification is performed using Long Short-Term Memory (LSTM) network with different structures and variants. The experimental results show that LSTM network outperforms traditional machine learning methods.
13 March 2025 in IEEE Sensors Journal
Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder characterized by difficulties in social interaction, communication, and repetitive behaviors. Diagnosing and intervening in ASD is challenging, yet early intervention is crucial for the development of affected children. We collected data from 261 participants, including 140 typically developing (TD) children and 121 ASD children, who participated in three experimental paradigms: human/object preference, joint attention, and sound response. Using HRNet, we extracted skeletal keypoint coordinates from video clips and applied the PoTion algorithm to generate joint motion trajectory maps, further analyzing the body movement characteristics of ASD and TD children. The study revealed significant differences in the body movement patterns of ASD children, characterized by a wider range of motion and irregular movement patterns. Additionally, by classifying the motion trajectory maps using a CNN network, we found that the joint attention paradigm achieved the highest classification accuracy at 87%, indicating that it contains more distinguishing information. This study proposed a framework based on MTCNN and HRNet algorithms for accurately extracting children’s body movement features in multi-person scenarios. The introduction of the PoTion algorithm for visualizing joint motion trajectories provided a more intuitive and precise method for analyzing the body movement characteristics of ASD and TD children. Our indepth analysis revealed significant differences in joint distances, angles, and velocities, offering new insights into the relationship between cognition and movement in ASD children.
8 July 2023 in Mobile Networks and Applications