Detection of ASD Children through Deep-Learning Application of fMRI
Abstract
:1. Introduction
- This study introduced a custom deep CNN for fMRI data, achieving high accuracy in distinguishing pediatric ASD from typical development.
- The study uncovered the model’s feature learning through in-depth analysis and feature-map visualization, enabling robust ASD discrimination.
- The investigation yielded invaluable insights via feature-map analysis, corroborating the efficacy of the CNN architecture in hierarchical feature learning.
2. Materials and Methodology
2.1. Data Collection and Preprocessing
2.2. Convolutional Neural Network Framework
2.3. Workflow for ASD Recognition Using CNNs
Algorithm 1: CNN Model for ASD Recognition |
Data: Preprocessed 4D fMRI data F(x,y,z,t) Result: ASD Recognition Performance 1 Step 1: Data Preprocessing and Slicing 2 for ti in T (Temporal Instances) do 3 V (x,y,z) = F(x,y,z,ti); 4 D = [S(x,z),C(y,z),A(x,y)]; 5 Add D to Dataset 6 Step 2: Dataset Partitioning 7 Divide Dataset into Training, Validation, and Test Sets 8 Step 3: Model Initialization 9 Initialize CNN Model Parameters θ 10 Step 4: Model Training for i = 1 to N (Iterations) do 11 Train CNN Model on Training Set using θ and Adam Optimizer; 12 Validate Model on Validation Set if Validation Loss ≤ Threshold 13 then 14 Break 15 Step 5: Model Testing 16 Test Trained Model on Test Set 17 Step 6: Performance Evaluation 18 Compute ASD Recognition Performance Metrics |
3. Experimental Setup
3.1. Parameters for CNN
3.2. Training Model
3.3. Evaluation Metrics
4. Results and Analysis
4.1. Model Performance
4.2. Feature Map Analysis
4.3. Discussion
5. Conclusions
- (1)
- The proffered CNN model exhibited extraordinary proficiency in differentiating ASD from typically developing (TD) subjects, attaining an accuracy metric of 99.39%, thereby markedly eclipsing prior scholarly endeavors. This underscored the model’s efficacy in discerning discriminative features within the fMRI datasets.
- (2)
- Scrutiny of the model’s feature maps corroborated its capabilities in hierarchical feature extraction, with the more advanced layers serving a pivotal role in demarcating ASD from TD. This suggested that the model was adept at learning increasingly intricate and abstract data representations.
- (3)
- The model furnished a rapid, unerring, and highly precise diagnostic apparatus for ASD screening and identification, with the potential to revolutionize conventional, subjective, and protracted diagnostic frameworks.
- (4)
- The model’s exceptional performance across a gamut of computational methodologies accentuated the efficacy of its unique architectural design and feature extraction paradigms for this specific classification task.
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Acknowledgments
Conflicts of Interest
References
- Kim, H.; Lim, C.-S.; Kaang, B.-K. Neuronal Mechanisms and Circuits Underlying Repetitive Behaviors in Mouse Models of Autism Spectrum Disorder. Behav. Brain Funct. 2016, 12, 3. [Google Scholar] [CrossRef] [PubMed]
- Al Sagheer, T.; Haida, O.; Balbous, A.; Francheteau, M.; Matas, E.; Fernagut, P.-O.; Jaber, M. Motor Impairments Correlate with Social Deficits and Restricted Neuronal Loss in an Environmental Model of Autism. Int. J. Neuropsychopharmacol. 2018, 21, 871–882. [Google Scholar] [CrossRef] [PubMed]
- Owens, D.G.C.; Johnstone, E.C. Precursors and Prodromata of Schizophrenia: Findings from the Edinburgh High Risk Study and Their Literature Context. Psychol. Med. 2006, 36, 1501–1514. [Google Scholar] [CrossRef] [PubMed]
- Martinussen, M.; Friborg, O.; Schmierer, P.; Kaiser, S.; Øvergård, K.T.; Neunhoeffer, A.-L.; Martinsen, E.W.; Rosenvinge, J.H. The Comorbidity of Personality Disorders in Eating Disorders: A Meta-Analysis. Eat. Weight. Disord. 2017, 22, 201–209. [Google Scholar] [CrossRef]
- Kaiser, M.D.; Hudac, C.M.; Shultz, S.; Lee, S.M.; Cheung, C.; Berken, A.M.; Deen, B.; Pitskel, N.B.; Sugrue, D.R.; Voos, A.C.; et al. Neural Signatures of Autism. Proc. Natl. Acad. Sci. USA 2010, 107, 21223–21228. [Google Scholar] [CrossRef]
- Liu, S.; Liang, B.; Wang, S.; Li, B.; Pan, L.; Wang, S. NF-GAT: A Node Feature-Based Graph Attention Network for ASD Classification. IEEE Open J. Eng. Med. Biol. 2023, 1–6. [Google Scholar] [CrossRef]
- Singh, S.; Malhotra, D.; Mengi, M. TransLearning ASD: Detection of Autism Spectrum Disorder Using Domain Adaptation and Transfer Learning-Based Approach on RS-FMRI Data. Artif. Intell. Commun. Technol. 2023, 863–871. [Google Scholar] [CrossRef]
- El Nakieb, Y.; Ali, M.T.; Elnakib, A.; Shalaby, A.; Mahmoud, A.; Soliman, A.; Barnes, G.N.; El-Baz, A. Understanding the Role of Connectivity Dynamics of Resting-State Functional MRI in the Diagnosis of Autism Spectrum Disorder: A Comprehensive Study. Bioengineering 2023, 10, 56. [Google Scholar] [CrossRef]
- Anderson, J.L.; Nielsen, J.A.; Froehlich, A.L.; DuBray, M.B.; Druzgal, T.J.; Cariello, A.N.; Cooperrider, J.R.; Zielinski, B.A.; Ravichandran, C.; Fletcher, P.T.; et al. Functional Connectivity Magnetic Resonance Imaging Classification of Autism. Brain 2011, 134, 3742–3754. [Google Scholar] [CrossRef]
- Polyakova, S. The Pathophysiological Rationale for Personalized Metabolic Therapy of Asd. Promising Treatments. Autism Dev. Disord. 2019, 17, 55–70. [Google Scholar] [CrossRef]
- Zhao, Y.; Dong, Q.; Zhang, S.; Zhang, W.; Chen, H.; Jiang, X.; Guo, L.; Hu, X.; Han, J.; Liu, T. Automatic Recognition of fMRI-Derived Functional Networks Using 3-D Convolutional Neural Networks. IEEE Trans. Biomed. Eng. 2018, 65, 1975–1984. [Google Scholar] [CrossRef]
- Hao, X.; An, Q.; Li, J.; Min, H.; Guo, Y.; Yu, M.; Qin, J. Exploring High-Order Correlations with Deep-Broad Learning for Autism Spectrum Disorder Diagnosis. Front. Neurosci. 2022, 16, 1046268. [Google Scholar] [CrossRef]
- Anirudh, R.; Thiagarajan, J.J. Bootstrapping Graph Convolutional Neural Networks for Autism Spectrum Disorder Classification. In Proceedings of the ICASSP 2019—2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Brighton, UK, 12–17 May 2019; IEEE: New York, NY, USA, 2019; pp. 3197–3201. [Google Scholar]
- Yin, W.; Mostafa, S.; Wu, F.-X. Diagnosis of Autism Spectrum Disorder Based on Functional Brain Networks with Deep Learning. J. Comput. Biol. 2021, 28, 146–165. [Google Scholar] [CrossRef] [PubMed]
- Jiao, Z.; Li, H.; Fan, Y. Improving Diagnosis of Autism Spectrum Disorder and Disentangling Its Heterogeneous Functional Connectivity Patterns Using Capsule Networks. In Proceedings of the 2020 IEEE 17th International Symposium on Biomedical Imaging (ISBI), Iowa City, IA, USA, 3–7 April 2020; IEEE: New York, NY, USA, 2020; pp. 1331–1334. [Google Scholar]
- Byeon, K.; Kwon, J.; Hong, J.; Park, H. Artificial Neural Network Inspired by Neuroimaging Connectivity: Application in Autism Spectrum Disorder. In Proceedings of the 2020 IEEE International Conference on Big Data and Smart Computing (BigComp), Busan, Republic of Korea, 13–16 February 2020; IEEE: New York, NY, USA, 2020; pp. 575–578. [Google Scholar]
- Di Martino, A.; Yan, C.-G.; Li, Q.; Denio, E.; Castellanos, F.X.; Alaerts, K.; Anderson, J.S.; Assaf, M.; Bookheimer, S.Y.; Dapretto, M.; et al. The Autism Brain Imaging Data Exchange: Towards a Large-Scale Evaluation of the Intrinsic Brain Architecture in Autism. Mol. Psychiatry 2013, 19, 659–667. [Google Scholar] [CrossRef] [PubMed]
- Jeyapratap, A.S.; Parker, D.; Yerys, B.E.; Verma, R.; Osmanlioglu, Y. Analysis of Structural Network Topology Alterations in Autism and Its Relationship with Autism Severity. bioRxiv 2022, 1–4. [Google Scholar] [CrossRef]
- Dong, X.L.; Rekatsinas, T. Data Integration and Machine Learning. Proc. Vldb Endow. 2018, 11, 2094–2097. [Google Scholar] [CrossRef]
- Rashid, M.; Singh, H.; Goyal, V. The Use of Machine Learning and Deep Learning Algorithms in Functional Magnetic Resonance Imaging—A Systematic Review. Expert Syst. 2020, 37, e12644. [Google Scholar] [CrossRef]
- Chawla, P.; Rana, S.B.; Kaur, H.; Singh, K. Computer-Aided Diagnosis of Autism Spectrum Disorder From EEG Signals Using Deep Learning With FAWT and Multiscale Permutation Entropy Features. Proc. Inst. Mech. Eng. Part H J. Eng. Med. 2022, 237, 282–294. [Google Scholar] [CrossRef] [PubMed]
- Bakkour, A.; Lewis-Peacock, J.A.; Poldrack, R.A.; Schonberg, T. Neural Mechanisms of Cue-Approach Training. Neuroimage 2017, 151, 92–104. [Google Scholar] [CrossRef]
- Bryan, C.J.; Rudd, M.D. The Importance of Temporal Dynamics in the Transition From Suicidal Thought to Behavior. Clinical Clin. Psychol. Sci. Pract. 2016, 23, 21–25. [Google Scholar] [CrossRef]
- Litjens, G.; Kooi, T.; Bejnordi, B.E.; Setio, A.A.A.; Ciompi, F.; Ghafoorian, M.; van der Laak, J.A.W.M.; van Ginneken, B.; Sánchez, C.I. A Survey on Deep Learning in Medical Image Analysis. Med. Image Anal. 2017, 42, 60–88. [Google Scholar] [CrossRef] [PubMed]
- Yang, Y.; Ge, Y.; Guo, L.; Wu, Q.; Peng, L.; Zhang, E.; Xie, J.; Li, Y.; Lin, T. Development and Validation of Two Artificial Intelligence Models for Diagnosing Benign, Pigmented Facial Skin Lesions. Ski. Res. Technol. 2020, 27, 74–79. [Google Scholar] [CrossRef] [PubMed]
- Sarvamangala, D.; Kulkarni, R.V. Convolutional Neural Networks in Medical Image Understanding: A Survey. Evol. Intell. 2021, 15, 1–22. [Google Scholar] [CrossRef] [PubMed]
- Yamins, D.L.K.; Hong, H.; Cadieu, C.F.; Solomon, E.A.; Seibert, D.; DiCarlo, J.J. Performance-Optimized Hierarchical Models Predict Neural Responses in Higher Visual Cortex. Proc. Natl. Acad. Sci. USA 2014, 111, 8619–8624. [Google Scholar] [CrossRef]
- Xu, J.; Yu, X. Detection of Concrete Structural Defects Using Impact Echo Based on Deep Networks. J. Test. Eval. 2021, 49, 20190801. [Google Scholar] [CrossRef]
- Xu, J.; Yu, X. Pavement Roughness Grade Recognition Based on One-Dimensional Residual Convolutional Neural Network. Sensors 2023, 23, 2271. [Google Scholar] [CrossRef]
- Szegedy, C.; Liu, W.; Jia, Y.; Sermanet, P.; Reed, S.; Anguelov, D.; Erhan, D.; Vanhoucke, V.; Rabinovich, A. Going Deeper with Convolutions. In Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA, 7–12 June 2015; IEEE: New York, NY, USA, 2015; pp. 1–9. [Google Scholar]
- Huang, G.; Liu, Z.; Maaten, L.V.D.; Weinberger, K.Q. Densely Connected Convolutional Networks. In Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA, 21–26 July 2017; pp. 2261–2269. [Google Scholar] [CrossRef]
- Ioffe, S.; Szegedy, C. Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. In Proceedings of the International Conference on Machine Learning, Lille, France, 6–11 July 2015; pp. 448–456. [Google Scholar]
- Srivastava, N.; Hinton, G.; Krizhevsky, A.; Sutskever, I.; Salakhutdinov, R. Dropout: A Simple Way to Prevent Neural Networks from Overfitting. J. Mach. Learn. Res. 2014, 15, 1929–1958. [Google Scholar]
- Hu, K.; Zhang, Z.; Niu, X.; Zhang, Y.; Cao, C.; Xiao, F.; Gao, X. Retinal Vessel Segmentation of Color Fundus Images Using Multiscale Convolutional Neural Network with an Improved Cross-Entropy Loss Function. Neurocomputing 2018, 309, 179–191. [Google Scholar] [CrossRef]
- Kingma, D.P.; Ba, J. Adam: A Method for Stochastic Optimization. arXiv 2014, arXiv:1412.6980. [Google Scholar] [CrossRef]
- Simonyan, K. Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv 2014, arXiv:1409.1556. [Google Scholar] [CrossRef]
- Lecun, Y.; Bottou, L.; Bengio, Y.; Haffner, P. Gradient-Based Learning Applied to Document Recognition. Proc. IEEE 1998, 86, 2278–2324. [Google Scholar] [CrossRef]
- Yakolli, N.; Anusha, V.; Khan, A.A.; Shubhashree, A.; Chatterjee, S. Enhancing the Diagnosis of Autism Spectrum Disorder Using Phenotypic, Structural, and Functional MRI Data. Innov. Syst. Softw. Eng. 2023. in review. [Google Scholar]
- Jiang, W.; Liu, S.; Zhang, H.; Sun, X.; Wang, S.-H.; Zhao, J.; Yan, J. CNNG: A Convolutional Neural Networks With Gated Recurrent Units for Autism Spectrum Disorder Classification. Front. Aging Neurosci. 2022, 14, 948704. [Google Scholar] [CrossRef] [PubMed]
- Husna, R.N.S.; Syafeeza, A.R.; Hamid, N.A.; Wong, Y.C.; Raihan, R.A. Functional Magnetic Resonance Imaging for Autism Spectrum Disorder Detection Using Deep Learning. J. Teknol. 2021, 83, 45–52. [Google Scholar] [CrossRef]
- Shi, C.; Xin, X.; Zhang, J. Domain Adaptation Using a Three-Way Decision Improves the Identification of Autism Patients from Multisite fMRI Data. Brain Sci. 2021, 11, 603. [Google Scholar] [CrossRef] [PubMed]
- Niu, K.; Guo, J.; Pan, Y.; Gao, X.; Peng, X.; Li, N.; Li, H. Multichannel Deep Attention Neural Networks for the Classification of Autism Spectrum Disorder Using Neuroimaging and Personal Characteristic Data. Complexity 2020, 2020, 1357853. [Google Scholar] [CrossRef]
- Guo, X.; Dominick, K.C.; Minai, A.A.; Li, H.; Erickson, C.A.; Lu, L.J. Diagnosing Autism Spectrum Disorder from Brain Resting-State Functional Connectivity Patterns Using a Deep Neural Network with a Novel Feature Selection Method. Front. Neurosci. 2017, 11, 460. [Google Scholar] [CrossRef]
- Ktena, S.I.; Parisot, S.; Ferrante, E.; Rajchl, M.; Lee, M.; Glocker, B.; Rueckert, D. Distance Metric Learning Using Graph Convolutional Networks: Application to Functional Brain Networks. In Medical Image Computing and Computer Assisted Intervention—MICCAI 2017; Descoteaux, M., Maier-Hein, L., Franz, A., Jannin, P., Collins, D.L., Duchesne, S., Eds.; Lecture Notes in Computer Science; Springer International Publishing: Cham, Switzerland, 2017; Volume 10433, pp. 469–477. ISBN 978-3-319-66181-0. [Google Scholar]
- Cilia, F.; Carette, R.; Elbattah, M.; Dequen, G.; Guérin, J.-L.; Bosche, J.; Vandromme, L.; Le Driant, B. Computer-Aided Screening of Autism Spectrum Disorder: Eye-Tracking Study Using Data Visualization and Deep Learning. JMIR Hum. Factors 2021, 8, e27706. [Google Scholar] [CrossRef]
Category | Value |
---|---|
Total | 126 |
ASDs | 56 |
TDs | 70 |
ASD Males | 51 |
ASD Females | 5 |
TD Males | 51 |
TD Females | 19 |
Age Range of ASDs (Years) | 7.13–18.00 |
Age Range of TDs (Years) | 6.47–18.00 |
Average Age (SD) (Years) | 12.02 (2.97) |
Average ADOS scores of ASDs (SD) | 11.20 (4.27) |
Layer (Type) | Output Shape | Param |
---|---|---|
conv2d 0 | (None, 62, 62, 32) | 896 |
batch normalization | (None, 62, 62, 32) | 128 |
max pooling2d | (None, 31, 31, 32) | 0 |
conv2d 1 | (None, 29, 29, 64) | 18,496 |
batch normalization 1 | (None, 29, 29, 64) | 256 |
max pooling2d 1 | (None, 14, 14, 64) | 0 |
conv2d 2 | (None, 12, 12, 128) | 73,856 |
batch normalization 2 | (None, 12, 12, 128) | 512 |
max pooling2d 2 | (None, 6, 6, 128) | 0 |
flatten | (None, 4608) | 0 |
dropout | (None, 4608) | 0 |
dense | (None, 1024) | 4,719,616 |
batch normalization 3 | (None, 1024) | 4096 |
dropout 1 (Dropout) | (None, 1024) | 0 |
dense 1 (Dense) | (None, 256) | 262,400 |
batch normalization 4 | (None, 256) | 1024 |
dropout 2 (Dropout) | (None, 256) | 0 |
dense 2 (Dense) | (None, 64) | 16,448 |
batch normalization 5 | (None, 64) | 256 |
dropout 3 (Dropout) | (None, 64) | 0 |
dense 3 (Dense) | (None, 2) | 130 |
Metric | Accuracy (%) | Precision (%) | Recall (%) | F1-Score (%) |
---|---|---|---|---|
Value | 99.39 | 99.85 | 98.80 | 99.32 |
Fold Number | 1 | 2 | 3 | 4 | 5 |
---|---|---|---|---|---|
Accuracy (%) | 99.39 | 99.68 | 99.70 | 99.32 | 99.09 |
References | Method | Pattern | Purpose | Accuracy (%) |
---|---|---|---|---|
Yakolli et al. [38] | CNN | FC and structural | Classification | 88.0 |
Jiang et al. [39] | 3D CNN | FC | Classification | 72.46 |
Hao et al. [12] | Deep-broad learning | ROI | Classification | 71.8 |
Husna et al. [40] | CNN | FC | Classification | 87.0 |
Shi et al. [41] | Decision model | ROI | Classification | 75.41 |
Niu et al. [42] | DANN | FC/ROI | Classification | 73.2 |
Byeon et al. [16] | RNN | FC | Classification | 74.54 |
Jiao et al. [15] | CapsNets | FC | Classification | 71.0 |
Yin et al. [14] | DNN, AE | ROI | Classification | 79.20 |
Anirudh et al. [13] | GCNN | ROI | Classification | 70.86 |
Zhao et al. [11] | 3D CNN | ROI | Classification | 70.1 |
Guo et al. [43] | DNN | FC | Classification | 86.36 |
Ktena et al. [44] | GCNN | ROI | Classification | 62.9 |
The proposed model | CNN | FC | Classification | 99.39 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Feng, M.; Xu, J. Detection of ASD Children through Deep-Learning Application of fMRI. Children 2023, 10, 1654. https://doi.org/10.3390/children10101654
Feng M, Xu J. Detection of ASD Children through Deep-Learning Application of fMRI. Children. 2023; 10(10):1654. https://doi.org/10.3390/children10101654
Chicago/Turabian StyleFeng, Min, and Juncai Xu. 2023. "Detection of ASD Children through Deep-Learning Application of fMRI" Children 10, no. 10: 1654. https://doi.org/10.3390/children10101654