Meta-matching to translate phenotypic predictive models from big to small data on structural MRI
Poster No:
1375
Submission Type:
Abstract Submission
Authors:
Naren Wulan1,2,3, Lijun An1,2,3, Chen Zhang1,2,3, Ru Kong1,2,3, Pansheng Chen1,2,3, Danilo Bzdok4,5, Simon Eickhoff6,7, Avram Holmes8, B. T. Thomas Yeo1,2,3,9,10
Institutions:
1Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS, Singapore, Singapore, 2Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore, 3N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore, Singapore, Singapore, 4McConnell Brain Imaging Centre (BIC), Montreal Neurol, McGill Universityogical Institute (MNI), Montreal, QC, Canada, 5Mila – Quebec Artificial Intelligence Institute, Montreal, QC, Canada, 6Institute for Systems Neuroscience, Medical Faculty, Heinrich-Heine University Düsseldorf, Düsseldorf, Germany, 7Institute of Neuroscience and Medicine, Brain & Behavior (INM-7), Research Center Jülich, Jülich, Germany, 8Department of Psychiatry, Brain Health Institute, Rutgers University, Piscataway, NJ, 9Integrative Sciences and Engineering Programme (ISEP), National University of Singapore, Singapore, Singapore, 10Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, MA, United States
First Author:
Naren Wulan
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore
Co-Author(s):
Lijun An
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore
Chen Zhang
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore
Ruby Kong
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore
Pansheng Chen
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore
Danilo Bzdok
McConnell Brain Imaging Centre (BIC), Montreal Neurol, McGill Universityogical Institute (MNI)|Mila – Quebec Artificial Intelligence Institute
Montreal, QC, Canada|Montreal, QC, Canada
McConnell Brain Imaging Centre (BIC), Montreal Neurol, McGill Universityogical Institute (MNI)|Mila – Quebec Artificial Intelligence Institute
Montreal, QC, Canada|Montreal, QC, Canada
Simon Eickhoff
Institute for Systems Neuroscience, Medical Faculty, Heinrich-Heine University Düsseldorf|Institute of Neuroscience and Medicine, Brain & Behavior (INM-7), Research Center Jülich
Düsseldorf, Germany|Jülich, Germany
Institute for Systems Neuroscience, Medical Faculty, Heinrich-Heine University Düsseldorf|Institute of Neuroscience and Medicine, Brain & Behavior (INM-7), Research Center Jülich
Düsseldorf, Germany|Jülich, Germany
B. T. Thomas Yeo
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore|Integrative Sciences and Engineering Programme (ISEP), National University of Singapore|Martinos Center for Biomedical Imaging, Massachusetts General Hospital
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore|Singapore, Singapore|Charlestown, MA, United States
Centre for Sleep & Cognition & Centre for Translational Magnetic Resonance Research, NUS|Department of Electrical and Computer Engineering, National University of Singapore|N.1 Institute for Health & Institute for Digital Medicine, National University of Singapore|Integrative Sciences and Engineering Programme (ISEP), National University of Singapore|Martinos Center for Biomedical Imaging, Massachusetts General Hospital
Singapore, Singapore|Singapore, Singapore|Singapore, Singapore|Singapore, Singapore|Charlestown, MA, United States
Introduction:
A central goal in neuroscience is understanding how brain imaging is associated with behavior. Structural MRI has been used to make individualized predictions in a variety of neurological and psychiatric disorders (Arbabshirani et al., 2017; Bhagwat et al., 2019; Cohen et al., 2021) due to its good contrasts. However, the prediction performance is strongly limited by a small sample size for many current MRI studies (Kharabian Masouleh et al., 2019; Poldrack et al., 2020; He et al., 2020). By transferring knowledge from large-scale source datasets (e.g. UK Biobank) to small target datasets, meta-matching has greatly improved prediction performance in functional MRI (He et al., 2022). Here, we tailored meta-matching approaches to predict new phenotypes in small boutique datasets with structural MRI.
Methods:
Our study departed from the UK Biobank (N=36,461, 67 phenotypes; Alfaro-Almagro et al., 2018), HCP-YA (N=1017, 35 phenotypes; Van Essen et al., 2013), and HCP-Aging (N=656, 45 phenotypes; Bookheimer et al., 2019).
We transferred models pretrained from meta-training set to meta-test sets. For within UK Biobank analysis, we randomly split UK Biobank dataset into meta-training set (N=26573, 33 phenotypes) and meta-test set (N=9888, 34 phenotypes). There is no overlap between participants or phenotypes across meta-training and meta-test sets. On meta-test set, K participants (K-shot, where K had a value of 10, 20, 50, 100, and 200) were randomly selected to mimic traditional small sample size studies, while the remaining participants in the meta-test set served for evaluation. Each random K-shot split was repeated 100 times to ensure stability. For cross-dataset analysis, the UK Biobank served as a meta-training set, while HCP-YA and HCP-Aging served as meta-test sets separately.
We adopted meta-matching by pretraining a 3D CNN model (Peng et al., 2021) on the meta-training set structural brain imaging to improve phenotypic prediction performance on meta-test sets (Figure 1). For baseline approaches, we considered the elastic net and direct transfer learning algorithm (Figure 1). The input of elastic net was morphometric measures (volumes and thickness from cortical and/or subcortical ROIs by FreeSurfer); the input of deep learning approaches is T1 images affine transformed to MNI152 standard space.
We transferred models pretrained from meta-training set to meta-test sets. For within UK Biobank analysis, we randomly split UK Biobank dataset into meta-training set (N=26573, 33 phenotypes) and meta-test set (N=9888, 34 phenotypes). There is no overlap between participants or phenotypes across meta-training and meta-test sets. On meta-test set, K participants (K-shot, where K had a value of 10, 20, 50, 100, and 200) were randomly selected to mimic traditional small sample size studies, while the remaining participants in the meta-test set served for evaluation. Each random K-shot split was repeated 100 times to ensure stability. For cross-dataset analysis, the UK Biobank served as a meta-training set, while HCP-YA and HCP-Aging served as meta-test sets separately.
We adopted meta-matching by pretraining a 3D CNN model (Peng et al., 2021) on the meta-training set structural brain imaging to improve phenotypic prediction performance on meta-test sets (Figure 1). For baseline approaches, we considered the elastic net and direct transfer learning algorithm (Figure 1). The input of elastic net was morphometric measures (volumes and thickness from cortical and/or subcortical ROIs by FreeSurfer); the input of deep learning approaches is T1 images affine transformed to MNI152 standard space.
Results:
Figure 2 (A) shows the prediction accuracy (Pearson's correlation) across all test phenotypes on the UK Biobank meta-test set. The boxplots represent 100 repetitions for K-shot. We can observe that meta-matching-based approaches (meta-matching finetune and meta-matching stacking) can significantly outperform the Elastic net baseline and direct transfer learning methods (for every K number).
The previous experiment results (Figure 2 (A)) suggested that meta-matching-based methods can perform well when transferring within the same dataset (e.g. UK Biobank). To demonstrate the generalization ability of meta-matching-based methods. approaches are also applied to the meta-test set in the HCP-YA dataset Figure 2 (B), and the HCP-Aging dataset Figure 2 (C) respectively. Figure 2 (B) and Figure 2 (C) show that the meta-matching-based methods can significantly outperform baseline methods in most cases when transferring from the meta-training dataset (UK Biobank) to the meta-test dataset (HCP-YA or HCP-Aging).
The previous experiment results (Figure 2 (A)) suggested that meta-matching-based methods can perform well when transferring within the same dataset (e.g. UK Biobank). To demonstrate the generalization ability of meta-matching-based methods. approaches are also applied to the meta-test set in the HCP-YA dataset Figure 2 (B), and the HCP-Aging dataset Figure 2 (C) respectively. Figure 2 (B) and Figure 2 (C) show that the meta-matching-based methods can significantly outperform baseline methods in most cases when transferring from the meta-training dataset (UK Biobank) to the meta-test dataset (HCP-YA or HCP-Aging).
Conclusions:
We adopted meta-matching from functional to structural imaging and achieved superior performance over elastic net and direct transfer learning on small datasets including HCP-YA and HCP-Aging. Our results showed the great potential of meta-matching framework in structural MRI-based behavior predictions.
Modeling and Analysis Methods:
Classification and Predictive Modeling 1
Methods Development 2
Keywords:
Design and Analysis
Machine Learning
Modeling
MRI
STRUCTURAL MRI
1|2Indicates the priority used for review
Provide references using author date format
Bhagwat, N., Pipitone, J., Voineskos, A. N., Chakravarty, M. M., & Initiative, A. s. D. N. (2019). An artificial neural network model for clinical score prediction in Alzheimer disease using structural neuroimaging measures. Journal of Psychiatry and Neuroscience, 44(4), 246-260.
Cohen, S. E., Zantvoord, J. B., Wezenberg, B. N., Bockting, C. L., & van Wingen, G. A. (2021). Magnetic resonance imaging for individual prediction of treatment response in major depressive disorder: a systematic review and meta-analysis. Translational psychiatry, 11(1), 168.
Kharabian Masouleh, S., Eickhoff, S. B., Hoffstaedter, F., Genon, S., & Initiative, A. s. D. N. (2019). Empirical examination of the replicability of associations between brain structure and psychological variables. elife, 8, e43464.
Poldrack, R. A., Huckins, G., & Varoquaux, G. (2020). Establishment of best practices for evidence for prediction: a review. JAMA psychiatry, 77(5), 534-540.
He, T., Kong, R., Holmes, A. J., Nguyen, M., Sabuncu, M. R., Eickhoff, S. B., . . . Yeo, B. T. (2020). Deep neural networks and kernel regression achieve comparable accuracies for functional connectivity prediction of behavior and demographics. NeuroImage, 206, 116276.
He, T., An, L., Chen, P., Chen, J., Feng, J., Bzdok, D., . . . Yeo, B. T. (2022). Meta-matching as a simple framework to translate phenotypic predictive models from big to small data. Nature neuroscience, 25(6), 795-804.
Alfaro-Almagro, F., Jenkinson, M., Bangerter, N. K., Andersson, J. L., Griffanti, L., Douaud, G., . . . Vallee, E. (2018). Image processing and Quality Control for the first 10,000 brain imaging datasets from UK Biobank. NeuroImage, 166, 400-424.
Van Essen, D. C., Smith, S. M., Barch, D. M., Behrens, T. E., Yacoub, E., Ugurbil, K., & Consortium, W.-M. H. (2013). The WU-Minn human connectome project: an overview. NeuroImage, 80, 62-79.
Bookheimer, S. Y., Salat, D. H., Terpstra, M., Ances, B. M., Barch, D. M., Buckner, R. L., . . . Elam, J. S. (2019). The lifespan human connectome project in aging: an overview. NeuroImage, 185, 335-348.
Peng, H., Gong, W., Beckmann, C. F., Vedaldi, A., & Smith, S. M. (2021). Accurate brain age prediction with lightweight deep neural networks. Medical Image Analysis, 68, 101871.