Efficient Diagnosis of Autism Spectrum Disorder Using Optimized Machine Learning Models Based on Structural MRI

被引:1
|
作者
Bahathiq, Reem Ahmed [1 ]
Banjar, Haneen [1 ,2 ]
Jarraya, Salma Kammoun [1 ]
Bamaga, Ahmed K. [3 ,4 ]
Almoallim, Rahaf [5 ]
机构
[1] King Abdulaziz Univ, Fac Comp & Informat Technol, Dept Comp Sci, Jeddah 21589, Saudi Arabia
[2] King Abdulaziz Univ, Ctr Artificial Intelligence Precis Med, Jeddah 21589, Saudi Arabia
[3] King Abdulaziz Univ, Fac Med, Dept Pediat, Pediat Neurol Unit, Jeddah 21589, Saudi Arabia
[4] King Abdulaziz Univ Hosp, Dept Pediat, Jeddah 21589, Saudi Arabia
[5] King Faisal Specialists Hosp & Res Ctr, Dept Radiol, Jeddah 21589, Saudi Arabia
来源
APPLIED SCIENCES-BASEL | 2024年 / 14卷 / 02期
关键词
autism spectrum disorder; structural magnetic resonance imaging; machine learning; classification; feature selection; Boruta; grey wolf optimizer;
D O I
10.3390/app14020473
中图分类号
O6 [化学];
学科分类号
0703 ;
摘要
Autism spectrum disorder (ASD) affects approximately 1.4% of the population and imposes significant social and economic burdens. Because its etiology is unknown, effective diagnosis is challenging. Advancements in structural magnetic resonance imaging (sMRI) allow for the objective assessment of ASD by examining structural brain changes. Recently, machine learning (ML)-based diagnostic systems have emerged to expedite and enhance the diagnostic process. However, the expected success in ASD was not yet achieved. This study evaluates and compares the performance of seven optimized ML models to identify sMRI-based biomarkers for early and accurate detection of ASD in children aged 5 to 10 years. The effect of using hyperparameter tuning and feature selection techniques are investigated using two public datasets from Autism Brain Imaging Data Exchange Initiative. Furthermore, these models are tested on a local Saudi dataset to verify their generalizability. The integration of the grey wolf optimizer with a support vector machine achieved the best performance with an average accuracy of 71% (with further improvement to 71% after adding personal features) using 10-fold Cross-validation. The optimized models identified relevant biomarkers for diagnosis, lending credence to their truly generalizable nature and advancing scientific understanding of neurological changes in ASD.
引用
收藏
页数:29
相关论文
共 50 条
  • [1] An Efficient Autism Spectrum Disorder Classification in Different Age Groups using Machine Learning Models
    Subhash, Ambika Rani
    Motagi, Ashwin Kumar U.
    [J]. INTERNATIONAL JOURNAL OF ONLINE AND BIOMEDICAL ENGINEERING, 2024, 20 (09)
  • [2] Efficient Machine Learning Models for Early Stage Detection of Autism Spectrum Disorder
    Bala, Mousumi
    Ali, Mohammad Hanif
    Satu, Md Shahriare
    Hasan, Khondokar Fida
    Moni, Mohammad Ali
    [J]. ALGORITHMS, 2022, 15 (05)
  • [3] Autism Spectrum Disorder Diagnosis using Optimal Machine Learning Methods
    Alteneiji, Maitha Rashid
    Alqaydi, Layla Mohammed
    Tariq, Muhammad Usman
    [J]. INTERNATIONAL JOURNAL OF ADVANCED COMPUTER SCIENCE AND APPLICATIONS, 2020, 11 (09) : 252 - 260
  • [4] Identifying and Predicting Autism Spectrum Disorder Based on Multi-Site Structural MRI With Machine Learning
    Duan, YuMei
    Zhao, WeiDong
    Luo, Cheng
    Liu, XiaoJu
    Jiang, Hong
    Tang, YiQian
    Liu, Chang
    Yao, DeZhong
    [J]. FRONTIERS IN HUMAN NEUROSCIENCE, 2022, 15
  • [5] A systematic review of structural MRI biomarkers in autism spectrum disorder: A machine learning perspective
    Pagnozzi, Alex M.
    Conti, Eugenia
    Calderoni, Sara
    Fripp, Jurgen
    Rose, Stephen E.
    [J]. INTERNATIONAL JOURNAL OF DEVELOPMENTAL NEUROSCIENCE, 2018, 71 : 68 - 82
  • [6] Machine learning for autism spectrum disorder diagnosis using structural magnetic resonance imaging: Promising but challenging
    Bahathiq, Reem Ahmed
    Banjar, Haneen
    Bamaga, Ahmed K.
    Jarraya, Salma Kammoun
    [J]. FRONTIERS IN NEUROINFORMATICS, 2022, 16
  • [7] Diagnosis of autism spectrum disorder based on functional brain networks and machine learning
    Caroline L. Alves
    Thaise G. L. de O. Toutain
    Patricia de Carvalho Aguiar
    Aruane M. Pineda
    Kirstin Roster
    Christiane Thielemann
    Joel Augusto Moura Porto
    Francisco A. Rodrigues
    [J]. Scientific Reports, 13
  • [8] Diagnosis of autism spectrum disorder based on functional brain networks and machine learning
    Alves, Caroline L.
    Toutain, Thaise G. L. de O.
    Aguiar, Patricia de Carvalho
    Pineda, Aruane M.
    Roster, Kirstin
    Thielemann, Christiane
    Porto, Joel Augusto Moura
    Rodrigues, Francisco A.
    [J]. SCIENTIFIC REPORTS, 2023, 13 (01)
  • [9] Analysis of Autism Spectrum Disorder Prediction using various Machine Learning Models
    Kumaravel, V
    HelenPrabha, K.
    [J]. 2024 INTERNATIONAL CONFERENCE ON ADVANCES IN COMPUTING, COMMUNICATION AND APPLIED INFORMATICS, ACCAI 2024, 2024,
  • [10] Has machine learning enhanced the diagnosis of autism spectrum disorder?
    Shirwaikar R.D.
    Sarwari I.
    Najam M.
    Shama H.M.
    [J]. Critical Reviews in Biomedical Engineering, 2023, 51 (01) : 1 - 14