Early Prediction of Cognitive Deficit using Deep Learning and Brain Connectome

The high risk of neurodevelopmental impairments is a major concern for parents and clinicians caring for premature babies. Annually, approximately 22,000 very preterm infants (i.e. ≤32 weeks gestational age) in the United States develop cognitive deficits. Unfortunately, cognitive deficits cannot be reliably diagnosed until 3 to 5 years of age. There is a gap in our knowledge about early identification of infants at high-risk for cognitive deficits. This gap limits our ability to target early interventions to the highest risk infants during periods of optimal neuroplasticity to enhance their ability to reach their full intellectual potential.

The human brain is a highly interactive system that exhibits both structural and functional units/networks. Diffusion Magnetic Resonance Imaging (dMRI) and resting-state functional connectivity MRI (fcMRI) have made possible quantitative mapping of the connections within and between these networks. Scientists have dubbed this the brain connectome and it has opened a window for observing the human mind. Research supports the notion that cognitive deficits may result from a disturbance/breakdown in the connectome and that the brain connectome shows high variability among subjects. What is not yet known, is whether this variability can enable personalized clinical prediction of neurodevelopment. We propose to develop a robust machine learning framework that can analyze integrated structural and functional brain connectome data obtained at term corrected age to make predictions about cognitive outcomes at 2 years of age in very preterm infants.

We have evaluated our models using a very preterm infant cohort. Please contact the PI, Dr. Lili He at Lili.He@cchmc.org for further data inquiries.


Softwares:

He L, Li H, Chen M, Wang J, Altaye M, Dillman JR, Parikh NA. (2022). Deep multimodal learning from MRI and clinical data for early prediction of neurodevelopmental deficits in very preterm infants. Front Neurosci. 15:753033. doi:10.3389/fnins.2021.753033. PMCID: PMC8525883. PMC Journal

Zhang H, Li H, Dillman, JR, Parikh, NA, He L. (2022). Multi-contrast MRI image synthesis using switchable cycle-consistent generative adversarial networks. Diagnostics. 12(4), 816. doi: 10.3390/diagnostics12040816. Journal Github

Chen M, Li H, Fan H, Dillman JR, Wang H, Altaye M, Zhang B, Parikh NA, He L. (2022). ConCeptCNN: A novel multi-filter convolutional neural network for the prediction of neurodevelopmental disorders using brain connectome. Medical Physics. doi: 10.1002/mp.15545. PMID: 35246986. PubMed Journal Github

Li H, Chen M, Wang J, Illapani VSP, Parikh NA and He L. (2021). Automatic segmentation of diffuse white matter abnormality on T2-weighted brain MRI using deep learning in very preterm infants. Radiol Artif Intell. 3(3). doi: 10.1148/ryai.2021200166. Journal Github

Li H., He L., Dudley JA., Maloney TC., Somasundaram E., Brady SL., Parikh NA., Dillman JR., (2020). DeepLiverNet: a deep transfer learning model for classifying liver stiffness using clinical and T2-weighted magnetic resonance imaging data in children and young adults. Pediatr Radiol. 2020 Oct 13. doi: 10.1007/s00247-020-04854-3. Epub ahead of print. PMID: 33048183. PubMed Journal Github

He L, Li H, Wang J, Chen M, Gozdas E, Dillman JR, Parikh NA. A multi-task, multi-stage deep transfer learning model for early prediction of neurodevelopment in very preterm infants. (2020). Sci Rep. 15;10(1):15072. PMID: 32934282; PMCID: PMC749223 PubMed Journal Github

Chen M, Li H, Wang J, Yuan W, Altaye M, Parikh NA, He L, (2020). Early prediction of cognitive deficit in very preterm infants using brain structural connectome with transfer learning enhanced deep convolutional neural networks, Front Neurosci, 18;14:858. PMID: 33041749; PMCID: PMC7530168. PubMed Journal Github

Chen M, Li H, Wang J, Dillman JR, Parikh NA, & He L. (2019). A multichannel deep neural network model analyzing multiscale functional brain connectome data for attention deficit hyperactivity disorder detection. Radiol Artif Intell. 11;2(1):e190012. doi: 10.1148/ryai.2019190012. PMID: 32076663; PMCID: PMC6996597. PMC Journal Github

Li H., Parikh N.A., Wang J., Merhar S., Chen M., Parikh M., Holland S., He L., (2019). Objective and automated detection of diffuse white matter abnormality in preterm infants using deep convolutional neural networks. Frontiers in Neuroscience, PMCID: PMC6591530 PMID: 31275101 PMC Journal Github

He L., Li H., Dudley JA., Maloney TC., Brady SL., Somasundaram E., Trout AT, Dillman JR., (2019). Machine Learning Prediction of Liver Stiffness Using Clinical and T2-Weighted MRI Radiomic Data. AJR Am J Roentgenol. 2019 Sep;213(3):592-601. doi: 10.2214/AJR.19.21082. Epub 2019 May 23. PMID: 31120779. PubMed Journal Github

Parikh MN., Li H., He L., (2019). Enhancing Diagnosis of Autism With Optimized Machine Learning Models and Personal Characteristic Data. Front Comput Neurosci. 2019 Feb 15;13:9. doi: 10.3389/fncom.2019.00009. PMID: 30828295; PMCID: PMC6384273. PubMed Journal Github

Li H., Parikh NA., He L., (2018). A Novel Transfer Learning Approach to Enhance Deep Neural Network Classification of Brain Functional Connectomes. Front Neurosci. 2018 Jul 24;12:491. doi: 10.3389/fnins.2018.00491. PMID: 30087587; PMCID: PMC6066582. PubMed Journal Github

He L., Li H., Holland SK., Yuan W., Altaye M., Parikh NA., (2018). Early prediction of cognitive deficits in very preterm infants using functional connectome data in an artificial neural network framework. Neuroimage Clin. 2018 Jan 31;18:290-297. doi: 10.1016/j.nicl.2018.01.032. PMID: 29876249; PMCID: PMC5987842. PubMed Journal Github


Installation:

The script code requires no installation. It has been tested with Python 3.7. Dependencies are listed as following.

Dependency:

  • Keras 2.2.4
  • Tensorflow 1.15
  • Numpy 1.17
  • Scikit-learn 0.21.3

Acknowledgement:

This study was supported by the National Institutes of Health grants R21-HD094085.

License:

Our script is released under MIT License.