Joint Learning of Multiple Differential Networks with fMRI data for Brain Connectivity Alteration Detection

In this study we focus on the problem of joint learning of multiple differential networks with function Magnetic Resonance Imaging (fMRI) data sets from multiple research centers. As the research centers may use different scanners and imaging parameters, joint learning of differential networks with fMRI data from different centers may reflect the underlying mechanism of neurological diseases from different perspectives while capturing the common structures. We transform the task as a penalized logistic regression problem, and exploit sparse group Minimax Concave Penalty (gMCP) to induce common structures among multiple differential networks and the sparse structures of each differential network. To further enhance the empirical performance, we develop an ensemble-learning procedure. We conduct thorough simulation study to assess the finite-sample performance of the proposed method and compare with state-of-the-art alternatives. We apply the proposed method to analyze fMRI datasets related with Attention Deficit Hyperactivity Disorder from various research centers. The identified common hub nodes and differential interaction patterns coincides with the existing experimental studies.. Keyword: Brain connectivity; Ensemble Learning; fMRI; Group minimax concave penalty; Network comparison; Logistic regression.

[1]  Ming Yuan,et al.  High Dimensional Inverse Covariance Matrix Estimation via Linear Programming , 2010, J. Mach. Learn. Res..

[2]  R. Tibshirani,et al.  Sparse inverse covariance estimation with the graphical lasso. , 2008, Biostatistics.

[3]  Pei Wang,et al.  Partial Correlation Estimation by Joint Sparse Regression Models , 2008, Journal of the American Statistical Association.

[4]  Lei Xie,et al.  A new insight into underlying disease mechanism through semi-parametric latent differential network model , 2018, bioRxiv.

[5]  Yang Feng,et al.  JDINAC: joint density-based non-parametric differential interaction network analysis and classification using high-dimensional sparse omics data , 2017, bioRxiv.

[6]  Hongzhe Li,et al.  Model selection and estimation in the matrix normal graphical model , 2012, J. Multivar. Anal..

[7]  Hong Yan,et al.  Incorporating prior information into differential network analysis using non‐paranormal graphical models , 2017, Bioinform..

[8]  Tianxi Cai,et al.  Testing Differential Networks with Applications to Detecting Gene-by-Gene Interactions. , 2015, Biometrika.

[9]  Jin Liu,et al.  Integrative analysis of multiple cancer genomic datasets under the heterogeneity model , 2013, Statistics in medicine.

[10]  Hong Yan,et al.  Joint Learning of Multiple Differential Networks With Latent Variables , 2019, IEEE Transactions on Cybernetics.

[11]  Kerstin Konrad,et al.  Is the ADHD brain wired differently? A review on structural and functional connectivity in attention deficit hyperactivity disorder , 2010, Human brain mapping.

[12]  Yumou Qiu,et al.  Inference on Multi-level Partial Correlations Based on Multi-subject Time Series Data , 2021, Journal of the American Statistical Association.

[13]  Shuheng Zhou Gemini: Graph estimation with matrix variate normal instances , 2012, 1209.5075.

[14]  S. Faraone,et al.  Meta-Analysis of Structural Imaging Findings in Attention-Deficit/Hyperactivity Disorder , 2007, Biological Psychiatry.

[15]  Chenlei Leng,et al.  Sparse Matrix Graphical Models , 2012 .

[16]  Lexin Li,et al.  Mixed-Effect Time-Varying Network Model and Application in Brain Connectivity Analysis , 2018, Journal of the American Statistical Association.

[17]  Concave group methods for variable selection and estimation in high-dimensional varying coefficient models , 2014 .

[18]  Yunzhang Zhu,et al.  Multiple matrix Gaussian graphs estimation , 2018, Journal of the Royal Statistical Society. Series B, Statistical methodology.

[19]  Ruibin Xi,et al.  Differential network analysis via lasso penalized D-trace loss , 2015, 1511.09188.

[20]  Cun-Hui Zhang Nearly unbiased variable selection under minimax concave penalty , 2010, 1002.4734.

[21]  Ramesh Srinivasan,et al.  Functional connectivity of frontal cortex in healthy and ADHD children reflected in EEG coherence. , 2007, Cerebral cortex.

[22]  Stephen M. Smith,et al.  The future of FMRI connectivity , 2012, NeuroImage.

[23]  Lexin Li,et al.  Matrix Graph Hypothesis Testing and Application in Brain Connectivity Alternation Detection , 2018 .

[24]  T. Cai,et al.  A Constrained ℓ1 Minimization Approach to Sparse Precision Matrix Estimation , 2011, 1102.2233.

[25]  Yong He,et al.  Simultaneous differential network analysis and classification for matrix-variate data with application to brain connectivity. , 2021, Biostatistics.

[26]  Jianqing Fan,et al.  Sure independence screening for ultrahigh dimensional feature space , 2006, math/0612857.

[27]  Yong He,et al.  BrainNet Viewer: A Network Visualization Tool for Human Brain Connectomics , 2013, PloS one.

[28]  V. Calhoun,et al.  Shared and distinct resting functional connectivity in children and adults with attention-deficit/hyperactivity disorder , 2020, Translational Psychiatry.

[29]  Yikai Wang,et al.  An Efficient and Reliable Statistical Method for Estimating Functional Connectivity in Large Scale Brain Networks Using Partial Correlation , 2016, Front. Neurosci..

[30]  Daniel S. Margulies,et al.  The Neuro Bureau ADHD-200 Preprocessed repository , 2016, NeuroImage.

[31]  T. Cai,et al.  Direct estimation of differential networks. , 2014, Biometrika.

[32]  Lei Xie,et al.  Brain connectivity alteration detection via matrix‐variate differential network model , 2020, Biometrics.

[33]  S. Rombouts,et al.  Structural and functional connectivity in children and adolescents with and without attention deficit/hyperactivity disorder , 2017, Journal of child psychology and psychiatry, and allied disciplines.

[34]  Lexin Li,et al.  Hypothesis testing of matrix graph model with application to brain connectivity analysis , 2015, Biometrics.

[35]  Wessel N. van Wieringen,et al.  Ridge estimation of inverse covariance matrices from high-dimensional data , 2014, Comput. Stat. Data Anal..

[36]  Quanquan Gu,et al.  Identifying gene regulatory network rewiring using latent differential graphical models , 2016, Nucleic acids research.

[37]  C. Vaidya,et al.  Neurodevelopmental abnormalities in ADHD. , 2012, Current topics in behavioral neurosciences.

[38]  Somnath Datta,et al.  Integrating gene regulatory pathways into differential network analysis of gene expression data , 2019, Scientific Reports.

[39]  Larry A. Wasserman,et al.  The huge Package for High-dimensional Undirected Graph Estimation in R , 2012, J. Mach. Learn. Res..

[40]  Catherine J. Stoodley The Cerebellum and Neurodevelopmental Disorders , 2015, The Cerebellum.

[41]  Genevera I. Allen,et al.  Mixed Effects Models for Resampled Network Statistics Improves Statistical Power to Find Differences in Multi-Subject Functional Connectivity , 2015, bioRxiv.