Using Artificial Neural Network to Detect Fetal Alcohol Spectrum Disorder in Children
Abstract
:1. Introduction
2. Related Work
2.1. Machine Learning
2.2. Computer-Aided Detection
3. Materials and Methods
3.1. Data
3.2. Data Analysis Using Machine Learning
3.2.1. Neural Network Configuration Model
3.2.2. Enhanced Neural Network Model
4. Results
4.1. Psychometric Data Classification
4.2. Antisaccade Data Classification
4.3. Prosaccade Task Data Classification
4.4. Memory-Guided Saccade Task Classification Results
4.5. Diffusion Tensor Imaging (DTI) Data Classification Results
5. Discussion
6. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Data Availability Statement
Acknowledgments
Conflicts of Interest
Appendix A
optimizer=Adam( ) loss_function= sparse_categorical_crossentropy num_classes=2 no_epoch=1000 verbosity = 0 #ANN model model = Sequential ( ) model.add (Dense (200, activation=LeakyReLU(alpha =0.2), input_shape = [20])) model.add (Dropout (0.2)) model.add (Dense (15, activation=LeakyReLU(alpha =0.2))) model.add (Dropout (0.2)) model.add (Dense (num_classes, activation= tf.nn.softmax)) #compile the model model.compile (opt imizer=optimizer , loss=loss_function , metrics=[’accuracy’]) # Fit data to model X_train=X.iloc[train ,:] y_train=y.iloc[train] history= model.fit(X_train ,y_train ,epochs=no_epoch, verbose=verbosity) # generate generalization metrics X_test = X.iloc[test , :] y_test = y.iloc[test] scores =model.evaluate(X_test,y_test,verbose=0) predict=model.predict(X_test) |
for header in [ ’ f1 ’ , ’ f2 ’ , ’ f3 ’ , ’ f4 ’ , ’ f5 ’ , ’ f6 ’ , ’ f7 ’ , ’ f8 ’ , ’ f9 ’ , ’ f10 ’ , ’ f11 ’ , ’ f12 ’ , ’ f13 ’ , ’ f14 ’ , ’ f15 ’ , ’ f16 ’ , ’ f17 ’ , ’ f18 ’ , ’ f19 ’ , ’ f20 ’ ] : feature_columns.append(feature_column.numeric_column(header)) feature_layer = tf.keras.layers.DenseFeatures(feature_columns) batch_size = 18 train_ds = df _ to_dataset(train , batch_size=batch_size) val_ds = df_to_dataset(val , shuffle=False , batch_size=batch_size) test_ds = df_to_dataset(test , shuffle=False , batch_ size=batch_size) train_ds model = tf.keras.Sequential([ feature_layer , layers.Dense(64, activation=’relu’), layers.Dropout(.1) , layers.Dense(128, activation=’sigmoid’), layers.Dropout(.1), layers.Dense(64, activation=’sigmoid’), layers.Dropout(.1), layers.Dense(128, activation=’relu’), layers.Dropout(.1), layers.Dense(1) ]) model.compile(optimizer=’adam’, loss=tf.keras.losses.BinaryCrossentropy(from_logits=True), metrics=[’accuracy’]) history=model.fit(train_ds,validation_data=val_ds,epochs=50) loss, accuracy = model.evaluate(test_ds) |
References
- O’Leary, C.; Bower, C. Guidelines for pregnancy: What’s an acceptable risk, and how is the evidence (finally) shaping up? Drug Alcohol. Rev. 2012, 31, 170–183. [Google Scholar] [CrossRef] [PubMed]
- McCallum, K.; Holland, K. ‘To drink or not to drink’: Media framing of evidence and debate about alcohol consumption in pregnancy. Crit. Public Health 2018, 28, 412–423. [Google Scholar] [CrossRef]
- Sun, C.; Shrivastava, A.; Singh, S.; Gupta, A. Revisiting Unreasonable Effectiveness of Data in Deep Learning Era. In Proceedings of the 2017 IEEE International Conference on Computer Vision (ICCV), Venice, Italy, 22–29 October 2017; pp. 843–852. [Google Scholar]
- Cook, J.L.; Green, C.R.; Lilley, C.M.; Anderson, S.M.; Baldwin, M.E.; Chudley, A.E.; Conry, J.L.; LeBlanc, N.; Loock, C.A.; Lutke, J.; et al. Fetal alcohol spectrum disorder: A guideline for diagnosis across the lifespan. CMAJ 2016, 188, 191–197. [Google Scholar] [CrossRef] [Green Version]
- Chudley, A.E. Diagnosis of fetal alcohol spectrum disorder: Current practices and future considerations. Biochem. Cell Biol. 2018, 96, 231–236. [Google Scholar] [CrossRef]
- Popova, S.; Lange, S.; Chudley, A.E.; Reynolds, J.N.; Rehm, J.; May, P.; Riley, E. World Health Organization International Study on the Prevalence of Fetal Alcohol Spectrum Disorder (FASD). Cent. Addit. Ment. Health. Available online: www.camh.ca (accessed on 1 May 2021).
- Popova, S.; Lange, S.; Probst, C.; Gmel, G.; Rehm, J. Estimation of national, regional, and global prevalence of alcohol use during pregnancy and fetal alcohol syndrome: A systematic review and meta-analysis. Lancet Glob. Health 2017, 5, e290–e299. [Google Scholar] [CrossRef] [Green Version]
- Paolozza, A.; Treit, S.; Beaulieu, C.; Reynolds, J.N. Diffusion tensor imaging of white matter and correlates to eye movement control and psychometric testing in children with prenatal alcohol exposure. Hum. Brain Mapp. 2016, 38, 444–456. [Google Scholar] [CrossRef] [PubMed]
- May, P.A.; Chambers, C.D.; Kalberg, W.O.; Zellner, J.; Feldman, H.; Buckley, D.; Kopald, D.; Hasken, J.M.; Xu, R.; Honerkamp-Smith, G.; et al. Prevalence of Fetal Alcohol Spectrum Disorders in 4 US Communities. JAMA 2018, 319, 474. [Google Scholar] [CrossRef] [PubMed]
- El Naqa, I.; Murphy, M.J. What is machine learning? In Machine Learning in Radiation Oncology; Springer: Berlin/Heidelberg, Germany, 2015; pp. 3–11. [Google Scholar]
- Mohri, M.; Rostamizadeh, A.; Talwalkar, A. Foundations of Machine Learning; MIT Press: Cambridge, MA, USA, 2018. [Google Scholar]
- Doi, K. Computer-aided diagnosis in medical imaging: Historical review, current status and future potential. Comput. Med. Imaging Graph. 2007, 31, 198–211. [Google Scholar] [CrossRef] [Green Version]
- Ronald, C. Computer Aided Detection (CAD): An Overview. Cancer Imaging Off. Publ. Int. Cancer Imaging Soc. 2005, 5, 17–19. [Google Scholar]
- Rodriguez, C.I.; Vergara, V.M.; Davies, S.; Calhoun, V.D.; Savage, D.D.; Hamilton, D.A. Detection of prenatal alcohol exposure using machine learning classification of resting-state functional network connectivity data. Alcohol 2021, 93, 25–34. [Google Scholar] [CrossRef]
- Zhang, C.; Paolozza, A.; Tseng, P.H.; Reynolds, J.N.; Munoz, D.P.; Itti, L. Detection of Children/Youth With Fetal Alcohol Spectrum Disorder Through Eye Movement, Psychometric, and Neuroimaging Data. Front. Neurol. 2019, 10. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Fang, S.; Liu, Y.; Huang, J.; Vinci-Booher, S.; Anthony, B.; Zhou, F. Facial Image Classification of Mouse Embryos for the Animal Model Study of Fetal Alcohol Syndrome. Proc. Symp. Appl. Comput. 2009, 2009, 852–856. [Google Scholar] [PubMed]
- Huang, J.; Jain, A.; Fang, S.; Riley, E. Using facial images to diagnose fetal alcohol syndrome (FAS). In Proceedings of the International Conference on Information Technology: Coding and Computing (ITCC’05)—Volume II; IEEE: Las Vegas, NV, USA, 2005; pp. 66–71. [Google Scholar]
- Kim, S.Y.; Moon, S.K.; Jung, D.C.; Hwang, S.I.; Sung, C.K.; Cho, J.Y.; Kim, S.H.; Lee, J.; Lee, H.J. Pre-operative prediction of advanced prostatic cancer using clinical decision support systems: Accuracy comparison between support vector machine and artificial neural network. Korean J. Radiol. 2011, 12, 588–594. [Google Scholar] [CrossRef]
- Li, Y.; Huang, C.; Ding, L.; Li, Z.; Pan, Y.; Gao, X. Deep learning in bioinformatics: Introduction, application, and perspective in the big data era. Methods 2019, 166, 4–21. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Ravi, D.; Wong, C.; Deligianni, F.; Berthelot, M.; Andreu-Perez, J.; Lo, B.; Yang, G.Z. Deep Learning for Health Informatics. IEEE J. Biomed. Health Inform. 2017, 21, 4–21. [Google Scholar] [CrossRef] [Green Version]
- Al-Shayea, Q.K. Artificial neural networks in medical diagnosis. Int. J. Comput. Sci. Issues 2011, 8, 150–154. [Google Scholar]
- Lundervold, A.S.; Lundervold, A. An overview of deep learning in medical imaging focusing on MRI. Z. Med. Phys. 2019, 29, 102–127. [Google Scholar] [CrossRef]
- Wozniak, J.R.; Muetzel, R.L. What does diffusion tensor imaging reveal about the brain and cognition in fetal alcohol spectrum disorders? Neuropsychol. Rev. 2011, 21, 133–147. [Google Scholar] [CrossRef]
- Little, G.; Beaulieu, C. Multivariate models of brain volume for identification of children and adolescents with fetal alcohol spectrum disorder. Hum. Brain Mapp. 2020, 41, 1181–1194. [Google Scholar] [CrossRef]
- Gao, Y.; Geras, K.J.; Lewin, A.A.; Moy, L. New Frontiers: An Update on Computer-Aided Diagnosis for Breast Imaging in the Age of Artificial Intelligence. AJR Am. J. Roentgenol. 2019, 212, 300–307. [Google Scholar] [CrossRef] [PubMed]
- Massari, L.; Bulletti, A.; Prasanna, S.; Mazzoni, M.; Frosini, F.; Vicari, E.; Pantano, M.; Staderini, F.; Ciuti, G.; Cianchi, F.; et al. A Mechatronic Platform for Computer Aided Detection of Nodules in Anatomopathological Analyses via Stiffness and Ultrasound Measurements. Sensors 2019, 19, 2512. [Google Scholar] [CrossRef] [Green Version]
- Niemeijer, M.; Loog, M.; Abràmoff, M.D.; Viergever, M.A.; Prokop, M.; van Ginneken, B. On Combining Computer-Aided Detection Systems. IEEE Trans. Med. Imaging 2011, 30, 215–223. [Google Scholar] [CrossRef] [PubMed]
- Nishikawa, R.M. Computer-aided Detection and Diagnosis. In Digital Mammography; Bick, U., Diekmann, F., Eds.; Springer: Berlin/Heidelberg, Germany, 2010; pp. 85–106. [Google Scholar]
- Cicerone, M.T.; Camp, C.H. Potential Roles for Spectroscopic Coherent Raman Imaging for Histopathology and Biomedicine. In Neurophotonics and Biomedical Spectroscopy; Alfano, R.R., Shi, L., Eds.; Elsevier: Amsterdam, The Netherlands, 2019; pp. 547–570. [Google Scholar]
- Ben-Cohen, A.; Greenspan, H. Liver lesion detection in CT using deep learning techniques. In Handbook of Medical Image Computing and Computer Assisted Intervention; Zhou, S.K., Rueckert, D., Fichtinger, G., Eds.; Academic Press: Cambridge, MA, USA, 2020; pp. 65–90. [Google Scholar]
- Katzen, J.; Dodelzon, K. A review of computer aided detection in mammography. Clin. Imaging 2018, 52, 305–309. [Google Scholar] [CrossRef]
- Suzuki, K. A supervised ’lesion-enhancement’ filter by use of a massive-training artificial neural network (MTANN) in computer-aided diagnosis (CAD). Phys. Med. Biol. 2019, 54, s31–s45. [Google Scholar] [CrossRef] [PubMed] [Green Version]
- Muhammad Ashfaq Khan, Y.K. Cardiac Arrhythmia Disease Classification Using LSTM Deep Learning Approach. Comput. Mater. Contin. 2021, 67, 427–443. [Google Scholar] [CrossRef]
- Valentine, M.; Bihm, D.C.; Wolf, L.; Hoyme, H.E.; May, P.A.; Buckley, D.; Kalberg, W.; Abdul-Rahman, O.A. Computer-Aided Recognition of Facial Attributes for Fetal Alcohol Spectrum Disorders. Pediatrics 2017, 140, e20162028. [Google Scholar] [CrossRef] [PubMed]
- Green, C.R.; Lebel, C.; Rasmussen, C.; Beaulieu, C.; Reynolds, J.N. Diffusion Tensor Imaging Correlates of Saccadic Reaction Time in Children with Fetal Alcohol Spectrum Disorder. Alcohol. Clin. Exp. Res. 2013, 37, 1499–1507. [Google Scholar] [CrossRef] [PubMed]
- Jones, D.K.; Leemans, A. Diffusion Tensor Imaging. In Magnetic Resonance Neuroimaging: Methods and Protocols; Methods in Molecular Biology; Modo, M., Bulte, J.W., Eds.; Humana Press: Totowa, NJ, USA, 2011; pp. 127–144. [Google Scholar]
- Hagmann, P.; Jonasson, L.; Maeder, P.; Thiran, J.P.; Wedeen, V.J.; Meuli, R. Understanding Diffusion MR Imaging Techniques: From Scalar Diffusion-weighted Imaging to Diffusion Tensor Imaging and Beyond. RadioGraphics 2006, 26, S205–S223. [Google Scholar] [CrossRef] [Green Version]
- Glass, L.; Moore, E.M.; Akshoomoff, N.; Jones, K.L.; Riley, E.P.; Mattson, S.N. Academic Difficulties in Children with Prenatal Alcohol Exposure: Presence, Profile, and Neural Correlates. Alcohol. Clin. Exp. Res. 2017, 41, 1024–1034. [Google Scholar] [CrossRef] [Green Version]
- Green, J.H. Fetal Alcohol Spectrum Disorders: Understanding the Effects of Prenatal Alcohol Exposure and Supporting Students. J. Sch. Health 2007, 77, 103–108. [Google Scholar] [CrossRef] [PubMed]
- Mohammad, S.; Page, S.J.; Wang, L.; Ishii, S.; Li, P.; Sasaki, T.; Basha, A.; Salzberg, A.; Quezado, Z.; Imamura, F.; et al. Kcnn2 blockade reverses learning deficits in a mouse model of fetal alcohol spectrum disorders. Nat. Neurosci. 2020, 23, 533–543. [Google Scholar] [CrossRef]
- Wozniak, J.R.; Muetzel, R.L.; Mueller, B.A.; McGee, C.L.; Freerks, M.A.; Ward, E.E.; Nelson, M.L.; Chang, P.N.; Lim, K.O. Microstructural Corpus Callosum Anomalies in Children With Prenatal Alcohol Exposure: An Extension of Previous Diffusion Tensor Imaging Findings. Alcohol. Clin. Exp. Res. 2009, 33, 1825–1835. [Google Scholar] [CrossRef]
- Rasmussen, C. Executive Functioning and Working Memory in Fetal Alcohol Spectrum Disorder. Alcohol. Clin. Exp. Res. 2005, 29, 1359–1367. [Google Scholar] [CrossRef] [Green Version]
- Sajda, P. Machine Learning for Detection and Diagnosis of Disease. Annu. Rev. Biomed. Eng. 2006, 8, 537–565. [Google Scholar] [CrossRef] [Green Version]
- Jo, T.; Nho, K.; Saykin, A.J. Deep Learning in Alzheimer’s Disease: Diagnostic Classification and Prognostic Prediction Using Neuroimaging Data. Front. Aging Neurosci. 2019, 11, 220. [Google Scholar] [CrossRef] [Green Version]
- Shukla, P.; Gupta, T.; Saini, A.; Singh, P.; Balasubramanian, R. A Deep Learning Frame-Work for Recognizing Developmental Disorders. In Proceedings of the 2017 IEEE Winter Conference on Applications of Computer Vision (WACV), Santa Rosa, CA, USA, 24–31 March 2017; pp. 705–714. [Google Scholar]
- Deng, L.; Hinton, G.; Kingsbury, B. New types of deep neural network learning for speech recognition and related applications: An overview. In Proceedings of the 2013 IEEE International Conference on Acoustics, Speech and Signal Processing, Vancouver, BC, Canada, 26–30 May 2013; pp. 8599–8603. [Google Scholar]
- Villada, F.; Muñoz, N.; García-Quintero, E. Redes Neuronales Artificiales aplicadas a la Predicción del Precio del Oro. Informa. Tecnol. 2016, 27, 143–150. [Google Scholar] [CrossRef]
- Ketkar, N. Deep Learning with Python: A Hands-On Introduction; Apress: New York, NY, USA, 2017. [Google Scholar]
- Stone, M. Cross-Validatory Choice and Assessment of Statistical Predictions. J. R. Stat. Soc. 1974, 36, 111–147. [Google Scholar] [CrossRef]
- Zhu, Y.; Li, C.; Luo, B.; Tang, J.; Wang, X. Dense Feature Aggregation and Pruning for RGBT Tracking. In Proceedings of the 27th ACM International Conference on Multimedia, MM ’19, Nice, France, 21–25 October 2019; Association for Computing Machinery: New York, NY, USA, 2019; pp. 465–472. [Google Scholar]
- Paneiro, G.; Rafael, M. Artificial neural network with a cross-validation approach to blast-induced ground vibration propagation modeling. Undergr. Space 2021, 6, 281–289. [Google Scholar] [CrossRef]
Test | Number of Features (Input Data) | Total Data | FASD | Control |
---|---|---|---|---|
Psychometric | 20 | 129 | 58 | 71 |
Antisaccade task | 15 | 174 | 68 | 106 |
Prosaccade task | 18 | 186 | 71 | 115 |
Memory-guide saccade task | 26 | 154 | 61 | 93 |
DTI | 48 | 76 | 41 | 35 |
Number of Neurons | Accuracy | K-Fold Cross-Validation | ||||
---|---|---|---|---|---|---|
IL | 1 HL | 2 HL | PT | PV | Avg (Accuracy) | Avg (Loss) |
20 | 15 | - | 80.72% | 57.00% | 71.22% (+– 11.23) | 1.07 |
25 | 15 | - | 80.72% | 55.00% | 72.69% (+– 12.22) | 0.92 |
25 | 20 | - | 90.24% | 75.55% | 77.37% (+– 11.09) | 0.98 |
25 | 30 | - | 85.00% | 65.63% | 69.87% (+– 9.19) | 1.68 |
25 | 20 | 15 | 88.00% | 65.63% | 72.82% (+– 10.01) | 1.09 |
50 | 15 | - | 93.98% | 60.00% | 66.54% (+– 14.99) | 1.97 |
100 | 50 | 25 | 91.00% | 55.00% | 73.65% (+– 10.39) | 2.21 |
200 | 15 | - | 97.59% | 64.00% | 69.04% (+– 11.29) | 1.79 |
200 | 50 | 50 | 97.00% | 64.00% | 69.74% (+– 8.79) | 2.78 |
Test | K-Fold Cross-Validation | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Method | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | Avg | |
Psycho | Accuracy (%) | 76.92 | 84.61 | 84.61 | 84.62 | 61.53 | 76.92 | 69.23 | 69.23 | 69.23 | 75 | 75.19% |
Loss | 0.56 | 0.42 | 0.45 | 0.50 | 1.29 | 0.47 | 0.76 | 0.61 | 0.58 | 0.36 | 0.60 | |
MAE | 0.23 | 0.15 | 0.15 | 0.15 | 0.38 | 0.23 | 0.31 | 0.31 | 0.31 | 0.25 | 0.24 | |
Antisacc | Accuracy (%) | 88.89 | 55.56 | 77.78 | 70.59 | 64.71 | 70.59 | 47.06 | 70.59 | 70.47 | 52.94 | 61.51% |
Loss | 0.40 | 0.86 | 0.46 | 0.72 | 0.69 | 0.70 | 0.73 | 0.51 | 0.64 | 0.70 | 0.64 | |
MAE | 0.11 | 0.44 | 0.22 | 0.29 | 0.35 | 0.29 | 0.53 | 0.29 | 0.24 | 0.47 | 0.32 | |
Prosacc | Accuracy (%) | 47.37 | 57.89 | 63.16 | 63.16 | 63.68 | 68.42 | 72.22 | 72.22 | 66.67 | 72.22 | 65.70% |
Loss | 0.70 | 1.30 | 0.66 | 1.59 | 0.63 | 0.60 | 0.58 | 0.59 | 0.66 | 0.60 | 0.79 | |
MAE | 0.53 | 0.42 | 0.37 | 0.37 | 0.26 | 0.32 | 0.28 | 0.28 | 0.33 | 0.28 | 0.34 | |
MG-sacc | Accuracy (%) | 68.75 | 56.25 | 68.75 | 81.25 | 66.67 | 73.33 | 60.00 | 66.67 | 26.67 | 53.33 | 62.16% |
Loss | 0.78 | 0.62 | 0.53 | 0.61 | 0.77 | 0.63 | 0.66 | 0.65 | 0.89 | 0.69 | 0.68 | |
MAE | 0.31 | 0.44 | 0.31 | 0.19 | 0.33 | 0.27 | 0.4 | 0.33 | 0.73 | 0.47 | 0.38 | |
DTI | Accuracy (%) | 56.25 | 53.33 | 66.67 | 80.00 | 60.00 | 63.25% | |||||
Loss | 0.80 | 0.67 | 0.55 | 0.82 | 0.77 | 0.72 | ||||||
MAE | 0.44 | 0.47 | 0.33 | 0.2 | 0.4 | 0.36 |
Test | ANN Accuracy | Zhang’s Model Accuracy | Difference |
---|---|---|---|
Psychometric | 88.46 | 78.26 | 10.2 |
Antisaccade | 74.07 | 65 | 9.07 |
Prosaccade | 72.24 | 61 | 11.24 |
Memory-guided saccade | 88.0 | 55 | 33 |
DTI | 75 | 67.39 | 7.61 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2021 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
Duarte, V.; Leger, P.; Contreras, S.; Fukuda, H. Using Artificial Neural Network to Detect Fetal Alcohol Spectrum Disorder in Children. Appl. Sci. 2021, 11, 5961. https://doi.org/10.3390/app11135961
Duarte V, Leger P, Contreras S, Fukuda H. Using Artificial Neural Network to Detect Fetal Alcohol Spectrum Disorder in Children. Applied Sciences. 2021; 11(13):5961. https://doi.org/10.3390/app11135961
Chicago/Turabian StyleDuarte, Vannessa, Paul Leger, Sergio Contreras, and Hiroaki Fukuda. 2021. "Using Artificial Neural Network to Detect Fetal Alcohol Spectrum Disorder in Children" Applied Sciences 11, no. 13: 5961. https://doi.org/10.3390/app11135961
APA StyleDuarte, V., Leger, P., Contreras, S., & Fukuda, H. (2021). Using Artificial Neural Network to Detect Fetal Alcohol Spectrum Disorder in Children. Applied Sciences, 11(13), 5961. https://doi.org/10.3390/app11135961