Content uploaded by Ming Li
Author content
All content in this area was uploaded by Ming Li on Jun 11, 2018
Content may be subject to copyright.
A preview of the PDF is not available
Identifying children with autism spectrum disorder based on their face processing abnormality: A machine learning framework
Autism Research
Abstract and Figures
The atypical face scanning patterns in individuals with Autism Spectrum Disorder (ASD) has been repeatedly discovered by previous research. The present study examined whether their face scanning patterns could be potentially useful to identify children with ASD by adopting the machine learning algorithm for the classification purpose. Particularly, we applied the machine learning method to analyze an eye movement dataset from a face recognition task [Yi et al., 2016], to classify children with and without ASD. We evaluated the performance of our model in terms of its accuracy, sensitivity, and specificity of classifying ASD. Results indicated promising evidence for applying the machine learning algorithm based on the face scanning patterns to identify children with ASD, with a maximum classification accuracy of 88.51%. Nevertheless, our study is still preliminary with some constraints that may apply in the clinical practice. Future research should shed light on further valuation of our method and contribute to the development of a multitask and multimodel approach to aid the process of early detection and diagnosis of ASD. Autism Res 2016. © 2016 International Society for Autism Research, Wiley Periodicals, Inc.
Figures - uploaded by Ming Li
Author content
All figure content in this area was uploaded by Ming Li
Content may be subject to copyright.
... Intricate gaze patterns linked to ASD characteristics were revealed by this methodology, which highlights the potential for early detection through AI-based diagnostics. In a study [45], researchers processed the eye movements of individuals during instances of eye contact. Their model accurately identified children with and without ASD by analyzing variations in their facial scanning patterns, reaching an 88.51% success rate. ...
Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder leading to an inability to socially communicate and in extreme cases individuals are completely dependent on caregivers. ASD detection at early ages is crucial as early detection can reduce the effect on social impairment. Deep learning models have shown capability to detect ASD earlier compared to traditional detection methods used by clinics and experts. Ensemble models, renowned for their ability to enhance predictive performance by combining multiple models, have emerged as a powerful tool in machine learning. This study harnesses the strength of ensemble learning to address the critical challenge of ASD diagnosis. This study proposed a deep ensemble model leveraging the strengths of VGG16 and Xception net trained on Facial Images for ASD detection overcoming limitations in existing datasets through extensive preprocessing. Proposed model preprocessed the training dataset of facial images by converting side posed images into frontal face images, using Histogram Equalization (HE) to enhance colors, data augmentation techniques application, and using the Hue Saturation Value (HSV) color model. By integrating the feature extraction strengths of VGG16 and Xception with fully connected layers, our model has achieved a notable 97% accuracy on the Kaggle ASD Face Image Dataset. This approach supports early detection of ASD and aligns with Sustainable Development Goal 3, which focuses on improving health and well-being.
... The classification achieved an accuracy of 74.22% [16]. A study [17] utilized eye movement patterns during face scanning to identify autism in children. The experiment involved 29 children with ASD and 29 typically developing (TD) children, aged 4 to 11 years, using a Tobii T60 eye tracker with a 60 Hz sample rate. ...
Autism Spectrum Disorder (ASD) is a lifelong neurodevelopmental disorder characterized by significant impairments in social interaction, communication, and cognitive functions, including linguistic and object recognition abilities. Early diagnosis is crucial as it can significantly improve an autistic child's social communication skills and overall quality of life. One of the characteristic hallmarks of ASD is the difficulty of making or maintaining eye contact. This paper uses Machine Learning (ML) and Deep Learning (DL) approaches to analyze ASD diagnosis through eye-tracking data. This work uses Artificial Neural Networks (ANN), Decision Tree (DT), Support Vector Machine (SVM), and K-Nearest Neighbor (KNN) to differentiate between ASD and Typically Developing (TD) individuals. This study achieved a notable accuracy of 98% using the Decision Tree (DT) model, marking a significant improvement over previous works. Other models, such as SVM, KNN, and ANN, achieved accuracies of 93.37%, 92.46%, and 73%, respectively.
Traditional medical diagnostic methods face bottlenecks such as high cost, poor accessibility, and delayed diagnosis in genetic syndromes, neurological disorders, psychiatric disorders, and endocrine disorders. Face-based machine learning (ML) technology provides a new path for early screening of diseases by analyzing facial phenotypes, dynamic expressions, facial skin, and 3D structural abnormalities, and is gradually becoming a clinically assisted screening tool. This paper provides a comprehensive overview of the applications, advances, and challenges of the technology. We summarize the range of diseases for which facial diagnosis is applicable and describe the basic process and related techniques for face-based ML diagnostic systems. In addition, this paper organizes the resources of current publicly available facial medical datasets and clarifies their disease coverage and sample size. Finally, possible future solutions to challenges hindering widespread adoption in clinical practice such as data bias, privacy, interpretability, generalizability, clinical value, and resource constraints are discussed. This review aims to provide researchers with a comprehensive foundation that integrates clinical perspectives, technological insights, and practical resources, to facilitate the development and successful implementation of face-based ML diagnostics in real-world clinical practice.
Autism Spectrum Disorder (ASD) affects approximately of the global population and is characterized by difficulties in social communication and repetitive or obsessive behaviors. Early detection of autism is crucial, as it allows therapeutic interventions to be initiated earlier, significantly increasing the effectiveness of treatments. However, diagnosing ASD remains a challenge, as it is traditionally carried out through methods that are often subjective and based on interviews and clinical observations. With the advancement of computer vision and pattern recognition techniques, new possibilities are emerging to automate and enhance the detection of characteristics associated with ASD, particularly in the analysis of facial features. In this context, image-based computational approaches must address challenges such as low data availability, variability in image acquisition conditions, and high-dimensional feature representations generated by deep learning models. This study proposes a novel framework that integrates data augmentation, multi-filtering routines, histogram equalization, and a two-stage dimensionality reduction process to enrich the representation in pre-trained and frozen deep learning neural network models applied to image pattern recognition. The framework design is guided by practical needs specific to ASD detection scenarios: data augmentation aims to compensate for limited dataset sizes; image enhancement routines improve robustness to noise and lighting variability while potentially highlighting facial traits associated with ASD; feature scaling standardizes representations prior to classification; and dimensionality reduction compresses high-dimensional deep features while preserving discriminative power. The use of frozen pre-trained networks allows for a lightweight, deterministic pipeline without the need for fine-tuning. Experiments are conducted using eight pre-trained models on a well-established benchmark facial dataset in the literature, comprising samples of autistic and non-autistic individuals. The results show that the proposed framework improves classification accuracy by up to points when compared to baseline models using pre-trained networks without any preprocessing strategies - as evidenced by the ResNet-50 architecture, which increased from to . Moreover, Transformer-based models, such as ViTSwin, reached up to accuracy, highlighting the robustness of the proposed approach. These improvements were observed consistently across different network architectures and datasets, under varying data augmentation, filtering, and dimensionality reduction configurations. A systematic ablation study further confirms the individual and collective benefits of each component in the pipeline, reinforcing the contribution of the integrated approach. These findings suggest that the framework is a promising tool for the automated detection of autism, offering an efficient improvement in traditional deep learning-based approaches to assist in early and more accurate diagnosis.
Autism spectrum Syndrome (ASS) is an evolving disability caused by transformations in human brain. Patients having this disease have complications such as lacking of social communiqué and interaction, restricted interests to participate in activities. But for patients suffering from ASD, these physiognomies make life very complicated. Children with this, have divergent facial landmarks, from normal children. These landmarks can be identified using deep learning models which can be used to detect ASD. But they models should be able to extract and produce the proper patterns of the face features. Thus, for this task, two variants of CNN models namely Xception andVGG19 were used for this task. The dataset required for training and testing of the application was collected from the Kaggle platform and consists of 2,940 face images comprising equal number of healthy and autistic patient images. The results showed that Xception model has an accuracy of 86% while the VGG19 model gives an accuracy of 81%. The results indicated that the proposed system can be used for ASD.
In addition to core symptoms, i.e., social interaction and communication difficulties and restricted and repetitive behaviors, autism is also characterized by aspects of superior perception [1]. One well-replicated finding is that of superior performance in visual search tasks, in which participants have to indicate the presence of an odd-one-out element among a number of foils [2-5]. Whether these aspects of superior perception contribute to the emergence of core autism symptoms remains debated [4, 6]. Perceptual and social interaction atypicalities could reflect co-expressed but biologically independent pathologies, as suggested by a "fractionable" phenotype model of autism [7]. A developmental test of this hypothesis is now made possible by longitudinal cohorts of infants at high risk, such as of younger siblings of children with autism spectrum disorder (ASD). Around 20% of younger siblings are diagnosed with autism themselves [8], and up to another 30% manifest elevated levels of autism symptoms [9]. We used eye tracking to measure spontaneous orienting to letter targets (O, S, V, and +) presented among distractors (the letter X; Figure 1). At 9 and 15 months, emerging autism symptoms were assessed using the Autism Observation Scale for Infants (AOSI; [10]), and at 2 years of age, they were assessed using the Autism Diagnostic Observation Schedule (ADOS; [11]). Enhanced visual search performance at 9 months predicted a higher level of autism symptoms at 15 months and at 2 years. Infant perceptual atypicalities are thus intrinsically linked to the emerging autism phenotype.
Copyright © 2015 The Authors. Published by Elsevier Ltd.. All rights reserved.
Although the prevalence of autism spectrum disorder (ASD) has risen sharply in the last few years reaching 1 in 68, the average age of diagnosis in the United States remains close to 4-well past the developmental window when early intervention has the largest gains. This emphasizes the importance of developing accurate methods to detect risk faster than the current standards of care. In the present study, we used machine learning to evaluate one of the best and most widely used instruments for clinical assessment of ASD, the Autism Diagnostic Observation Schedule (ADOS) to test whether only a subset of behaviors can differentiate between children on and off the autism spectrum. ADOS relies on behavioral observation in a clinical setting and consists of four modules, with module 2 reserved for individuals with some vocabulary and module 3 for higher levels of cognitive functioning. We ran eight machine learning algorithms using stepwise backward feature selection on score sheets from modules 2 and 3 from 4540 individuals. We found that 9 of the 28 behaviors captured by items from module 2, and 12 of the 28 behaviors captured by module 3 are sufficient to detect ASD risk with 98.27% and 97.66% accuracy, respectively. A greater than 55% reduction in the number of behaviorals with negligible loss of accuracy across both modules suggests a role for computational and statistical methods to streamline ASD risk detection and screening. These results may help enable development of mobile and parent-directed methods for preliminary risk evaluation and/or clinical triage that reach a larger percentage of the population and help to lower the average age of detection and diagnosis.
In the present work, we have undertaken a proof-of-concept study to determine whether a simple upper-limb movement could be useful to accurately classify low-functioning children with autism spectrum disorder (ASD) aged 2-4. To answer this question, we developed a supervised machine-learning method to correctly discriminate 15 preschool children with ASD from 15 typically developing children by means of kinematic analysis of a simple reach-to-drop task. Our method reached a maximum classification accuracy of 96.7 % with seven features related to the goal-oriented part of the movement. These preliminary findings offer insight into a possible motor signature of ASD that may be potentially useful in identifying a well-defined subset of patients, reducing the clinical heterogeneity within the broad behavioral phenotype.
Current approaches for diagnosing autism have high diagnostic validity but are time consuming and can contribute to delays in arriving at an official diagnosis. In a pilot study, we used machine learning to derive a classifier that represented a 72% reduction in length from the gold-standard Autism Diagnostic Observation Schedule-Generic (ADOS-G), while retaining >97% statistical accuracy. The pilot study focused on a relatively small sample of children with and without autism. The present study sought to further test the accuracy of the classifier (termed the observation-based classifier (OBC)) on an independent sample of 2616 children scored using ADOS from five data repositories and including both spectrum (n=2333) and non-spectrum (n=283) individuals. We tested OBC outcomes against the outcomes provided by the original and current ADOS algorithms, the best estimate clinical diagnosis, and the comparison score severity metric associated with ADOS-2. The OBC was significantly correlated with the ADOS-G (r=-0.814) and ADOS-2 (r=-0.779) and exhibited >97% sensitivity and >77% specificity in comparison to both ADOS algorithm scores. The correspondence to the best estimate clinical diagnosis was also high (accuracy=96.8%), with sensitivity of 97.1% and specificity of 83.3%. The correlation between the OBC score and the comparison score was significant (r=-0.628), suggesting that the OBC provides both a classification as well as a measure of severity of the phenotype. These results further demonstrate the accuracy of the OBC and suggest that reductions in the process of detecting and monitoring autism are possible.
This study employed graph theory and machine learning analysis of multiparametric MRI data to improve characterization and prediction in autism spectrum disorders (ASD). Data from 127 children with ASD (13.5±6.0 years) and 153 age- and gender-matched typically developing children (14.5±5.7 years) were selected from the multi-center Functional Connectome Project. Regional gray matter volume and cortical thickness increased, whereas white matter volume decreased in ASD compared to controls. Small-world network analysis of quantitative MRI data demonstrated decreased global efficiency based on gray matter cortical thickness but not with functional connectivity MRI (fcMRI) or volumetry. An integrative model of 22 quantitative imaging features was used for classification and prediction of phenotypic features that included the autism diagnostic observation schedule, the revised autism diagnostic interview, and intelligence quotient scores. Among the 22 imaging features, four (caudate volume, caudate-cortical functional connectivity and inferior frontal gyrus functional connectivity) were found to be highly informative, markedly improving classification and prediction accuracy when compared with the single imaging features. This approach could potentially serve as a biomarker in prognosis, diagnosis, and monitoring disease progression.
Thesupport-vector network is a new learning machine for two-group classification problems. The machine conceptually implements the following idea: input vectors are non-linearly mapped to a very high-dimension feature space. In this feature space a linear decision surface is constructed. Special properties of the decision surface ensures high generalization ability of the learning machine. The idea behind the support-vector network was previously implemented for the restricted case where the training data can be separated without errors. We here extend this result to non-separable training data.High generalization ability of support-vector networks utilizing polynomial input transformations is demonstrated. We also compare the performance of the support-vector network to various classical learning algorithms that all took part in a benchmark study of Optical Character Recognition.
Previous work based on observations of home videotapes indicates that differences can be detected between infants with autism spectrum disorder and infants with typical development at I year of age, The present study addresses the question of whether autism can be distinguished from mental retardation by I year of age. Home videotapes of first birthday parties from 20 infants later diagnosed with autism spectrum disorder, 14 infants later diagnosed with mental retardation (without autism), and 20 typically developing infants were coded by blind raters with respect to the frequencies of specific social and communicative behaviors and repetitive motor actions. Results indicated that 1-year-olds with autism spectrum disorder can be distinguished from 1-year-olds with typical development and those with mental retardation. The infants with autism spectrum disorder looked at others and oriented to their names less frequently than infants with mental retardation. The infants with autism spectrum disorder and those with mental retardation used gestures and looked to objects held by others less frequently and engaged in repetitive motor actions more frequently than typically developing infants, These results indicate that autism can be distinguished from mental retardation and typical development by I year of age.
The support-vector network is a new learning machine for two-group classification problems. The machine conceptually implements the following idea: input vectors are non-linearly mapped to a very high-dimension feature space. In this feature space a linear decision surface is constructed. Special properties of the decision surface ensures high generalization ability of the learning machine. The idea behind the support-vector network was previously implemented for the restricted case where the training data can be separated without errors. We here extend this result to non-separable training data. High generalization ability of support-vector networks utilizing polynomial input transformations is demonstrated. We also compare the performance of the support-vector network to various classical learning algorithms that all took part in a benchmark study of Optical Character Recognition.