Detection of autism spectrum disorder (ASD) in children and adults using machine learning

Scientific Reports, Journal Year: 2023, Volume and Issue: 13(1)

Published: June 13, 2023

Autism spectrum disorder (ASD) presents a neurological and developmental that has an impact on the social cognitive skills of children causing repetitive behaviours, restricted interests, communication problems difficulty in interaction. Early diagnosis ASD can prevent from its severity prolonged effects. Federated learning (FL) is one most recent techniques be applied for accurate diagnoses early stages or prevention long-term In this article, FL technique been uniquely autism detection by training two different ML classifiers including logistic regression support vector machine locally classification factors adults. Due to FL, results obtained these have transmitted central server where meta classifier trained determine which approach Four patient datasets, each containing more than 600 records effected adults repository features extraction. The proposed model predicted with 98% accuracy (in children) 81% adults).

Language: Английский

---

Autism Spectrum Disorder detection framework for children based on federated learning integrated CNN-LSTM

Computers in Biology and Medicine, Journal Year: 2023, Volume and Issue: 166, P. 107539 - 107539

Published: Oct. 4, 2023

Language: Английский

---

A new diagnostic autism spectrum disorder (DASD) strategy using ensemble diagnosis methodology based on blood tests

Health Information Science and Systems, Journal Year: 2023, Volume and Issue: 11(1)

Published: Aug. 14, 2023

Autism Spectrum Disorder (ASD) is a complex neurodevelopmental disease that impacts child's way of behavior and social communication. In early childhood, children with ASD typically exhibit symptoms such as difficulty in interaction, limited interests, repetitive behavior. Although there are disease, most people do not understand these therefore have enough knowledge to determine whether or child has ASD. Thus, detection based on accurate diagnosis model Artificial Intelligence (AI) techniques critical process reduce the spread control it early. Through this paper, new Diagnostic (DASD) strategy presented quickly accurately detect children. DASD contains two layers called Data Filter Layer (DFL) (DL). Feature selection outlier rejection processes performed DFL filter dataset from less important features incorrect data before using diagnostic method DL diagnose patients. DFL, Binary Gray Wolf Optimization (BGWO) technique used select significant set while Genetic Algorithm (BGA) eliminate invalid training data. Then, Ensemble Diagnosis Methodology (EDM) precisely main contribution EDM consists several models including Enhanced K-Nearest Neighbors (EKNN) one them. EKNN represents hybrid consisting three methods (KNN), Naïve Bayes (NB), Chimp (COA). NB weighed convert feature space weight space. COA generation size dataset. Finally, KNN applied reduced small size. blood tests test proposed against other recent strategies [1]. It concluded superior many performance measures accuracy, error, recall, precision, micro_average macro_average F1-measure, implementation-time values equal 0.93, 0.07, 0.83, 0.82, 0.80, 0.79, 0.81, 1.5 s respectively.

Language: Английский

---

High-performance soil class delineation via UMAP coupled with machine learning in Kurdistan Province, Iran

Geoderma Regional, Journal Year: 2024, Volume and Issue: 36, P. e00754 - e00754

Published: Jan. 7, 2024

Language: Английский

---

Multimodal Deep Learning in Early Autism Detection—Recent Advances and Challenges

Published: Jan. 23, 2024

Autism spectrum disorder (ASD) is a global concern, with prevalence rate of approximately 1 in 36 children according to estimates from the Centers for Disease Control and Prevention (CDC). Diagnosing ASD poses challenges due absence definitive medical test. Instead, doctors rely on comprehensive evaluation child's developmental background behavior reach diagnosis. Although can occasionally be identified aged 18 months or younger, reliable diagnosis by an experienced professional typically made age two. Early detection crucial timely interventions improved outcomes. In recent years, field early has been greatly impacted emergence deep learning models, which have brought about revolution improving accuracy efficiency detection. The objective this review paper examine progress through utilization multimodal techniques. analysis revealed that integrating multiple modalities, including neuroimaging, genetics, behavioral data, key achieving higher It also evident that, while neuroimaging data holds promise potential contribute detection, it most effective when combined other modalities. Deep their ability analyze complex patterns extract meaningful features large datasets, offer great addressing challenge Among various models used, CNN, DNN, GCN, hybrid exhibited encouraging outcomes ASD. highlights significance developing accurate easily accessible tools utilize artificial intelligence (AI) aid healthcare professionals, parents, caregivers symptom recognition. These would enable interventions, ensuring necessary actions are taken during initial stages.

Language: Английский

---

Investigating Multiclass Autism Spectrum Disorder Classification using Machine Learning Techniques

e-Prime - Advances in Electrical Engineering Electronics and Energy, Journal Year: 2024, Volume and Issue: 8, P. 100602 - 100602

Published: May 18, 2024

The diagnosis and classification of autism spectrum disorder (ASD) presents anatomical difficulty owing to the existence a wide range symptoms that may be organized into many categories. present research investigates efficacy machine learning methods for facilitating recognition individuals who have been diagnosed with ASD. primary aim this study has assess effectiveness multiple algorithms based on in identifying intricate patterns seen datasets related ASD, which includes diagnostic results indicate Logistic Regression approach demonstrated great levels accuracy, rates 94.3% children 99% adolescents binary system. Similarly, it reported Support Vector Machine (SVM) had superior performance compared all other systems test focused adults exclusively, an accuracy rate 98.5%. Moreover, supplementary series experiments conducted combined dataset children, adolescents, resulted observation SVM exhibited notable 97.2% 99.55% multiclass classification, encompassing from diverse age groups. provide evidence favor progress achieved treatment ASD as result capacity detect categorize at earlier developmental phase.

Language: Английский

---

Computer-Aided Autism Spectrum Disorder Diagnosis With Behavior Signal Processing

IEEE Transactions on Affective Computing, Journal Year: 2023, Volume and Issue: 14(4), P. 2982 - 3000

Published: Jan. 23, 2023

Behavioral observation plays an essential role in the diagnosis of Autism Spectrum Disorder (ASD) by analyzing children's atypical patterns social activities (e.g., impaired interaction, restricted interests, and repetitive behavior). To date, this process still heavily relies on questionnaire survey, clinical observation, or retrospective video analysis, leading to high demand for professionals with massive labor costs. This article proposes a standardized platform stimulating, gathering, analyzing, modeling, interpreting human behavioral data application computer-aided ASD diagnosis. By structured assessment process, proposed system can automatically evaluate multiple interaction skills using captured audio-visual provide final diagnostic suggestions. We collect multimodal database 95 participants (71 children 24 age-matched typical controls) real clinic environment, Third Affiliated Hospital Sun Yat-sen University, China. On database, our obtains accuracy 88.42% identifying average age months, representing performance comparable top-level experts. As unified replicable solution, it has good potential be promoted less developed areas limited high-quality medical resources.

Language: Английский

---

Developing a Simplified Measure to Predict the Risk of Autism Spectrum Disorders: Abbreviating the M-CHAT-R Using a Machine Learning Approach in China

Psychiatry Research, Journal Year: 2025, Volume and Issue: 344, P. 116353 - 116353

Published: Jan. 5, 2025

Early screening for autism spectrum disorder (ASD) is crucial, yet current assessment tools in Chinese primary child care are limited efficacy. This study aims to employ machine learning algorithms identify key indicators from the 20-item Modified Checklist Autism Toddlers, revised (M-CHAT-R) combining with ASD-related sociodemographic and environmental factors, distinguish ASD typically developing children. Data our prior validation of M-CHAT-R (August 2016-March 2017, n = 6,049 toddlers) were reviewed. We extracted data integrated 17 risk factors associated development strengthen M-CHAT-R's screening. Five feature selection methods used extract subsets original set. Six applied optimal subset distinguishing clinically diagnosed toddlers toddlers. Nine features grouped into three subsets: 1 contained unanimously recommended items (A1 [Follows point], A3 [Pretend play], A9 [Brings objects show], A10 [Response name] A16 [Gazing following]). Subset 2 added two (A17 [Gaining parent's attention] A18 [Understands what said]), 3 included more (A8 [Interest other children] child's age). The top-performing algorithm resulted a seven-item classifier 92.5 % sensitivity, 90.1 specificity, 10.0 positive predictive value. Machine classifiers effectively differentiate using reduced item highlights clinical significance learning-optimized models health centers broader applications.

Language: Английский

---

Machine Learning Approaches for Autism Spectrum Disorder Detection: A Systematic Review of Age-Specific Applications and Performance Metrics

International Journal of Scientific Research in Science and Technology, Journal Year: 2025, Volume and Issue: 12(1), P. 213 - 227

Published: Jan. 27, 2025

Autism Spectrum Disorder is one of the biggest concerns in healthcare sector, and it’s crucial to diagnose it at an early stage for patients with Disorder. This review focuses on use machine learning diagnosing Disorder, drawing data from 100 papers between 2015 2024. We touched every possible method starting classic ones like Support Vector Machines (SVMs) new federated learning. Proving actually great since very precise (up 98%) while keeping people’s information personal, which a matter industry. But cannot write-off basic framework where people standard models such as SVMs, this point achieve around 92% accuracy. Also, they are more convenient be implemented small clinics that do not possess many computers, etcetera. suggests most suitable ML approaches detection need consider accuracy, privacy availability resources. Lately, developed technologies provide even better outcomes; nevertheless, conventional techniques terrific options without much complicated systems available. Thus, study offers meaningful suggestions facilitate choice methods based comparison these approaches. In sum, spans existing gap research advancements state-of-art practical settings provides important recommendations enhancing screening across various contexts.

Language: Английский

---

Innovative Deep Learning Approaches in Autism Research

Published: Jan. 1, 2025

Language: Английский

---