Small-sample brain mapping: sparse recovery on spatially correlated designs with randomization and clustering

Functional neuroimaging can measure the brain's response to an external stimulus. It is used to perform brain mapping: identifying from these observations the brain regions involved. This problem can be cast into a linear supervised learning task where the neuroimaging data are used as predictors for the stimulus. Brain mapping is then seen as a support recovery problem. On functional MRI (fMRI) data, this problem is particularly challenging as i) the number of samples is small due to limited acquisition time and ii) the variables are strongly correlated. We propose to overcome these dificulties using sparse regression models over new variables obtained by clustering of the original variables. The use of randomization techniques, e.g. bootstrap samples, and clustering of the variables improves the recovery properties of sparse methods. We demonstrate the benefit of our approach on an extensive simulation study as well as two fMRI datasets.

[1]  Martin J. Wainwright,et al.  Sharp Thresholds for High-Dimensional and Noisy Sparsity Recovery Using $\ell _{1}$ -Constrained Quadratic Programming (Lasso) , 2009, IEEE Transactions on Information Theory.

[2]  A. Ishai,et al.  Distributed and Overlapping Representations of Faces and Objects in Ventral Temporal Cortex , 2001, Science.

[3]  H. Zou,et al.  Regularization and variable selection via the elastic net , 2005 .

[4]  Bin Yu,et al.  On Model Selection Consistency of the Elastic Net When p >> n , 2008 .

[5]  Chih-Jen Lin,et al.  LIBLINEAR: A Library for Large Linear Classification , 2008, J. Mach. Learn. Res..

[6]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[7]  Gaël Varoquaux,et al.  Scikit-learn: Machine Learning in Python , 2011, J. Mach. Learn. Res..

[8]  Francis R. Bach,et al.  Self-concordant analysis for logistic regression , 2009, ArXiv.

[9]  Jean-Philippe Vert,et al.  Group lasso with overlap and graph lasso , 2009, ICML '09.

[10]  A. Ng.  Feature selection, L1 vs. L2 regularization, and rotational invariance , 2004, Twenty-first international conference on Machine learning - ICML '04.

[11]  A. Ravishankar Rao,et al.  Prediction and interpretation of distributed neural activity with sparse models , 2009, NeuroImage.

[12]  R. Tibshirani.  Regression Shrinkage and Selection via the Lasso , 1996 .

[13]  Daniel Müllner,et al.  Modern hierarchical, agglomerative clustering algorithms , 2011, ArXiv.

[14]  N. Meinshausen,et al.  Stability selection , 2008, 0809.2932.

[15]  Sabrina M. Tom,et al.  The Neural Basis of Loss Aversion in Decision-Making Under Risk , 2007, Science.

[16]  Tom M. Mitchell,et al.  Learning to Decode Cognitive States from Brain Images , 2004, Machine Learning.

[17]  E. Candès,et al.  Stable signal recovery from incomplete and inaccurate measurements , 2005, math/0503066.

[18]  Gaël Varoquaux,et al.  A supervised clustering approach for fMRI-based inference of brain states , 2011, Pattern Recognit..

[19]  Stéphane Mallat,et al.  Matching pursuits with time-frequency dictionaries , 1993, IEEE Trans. Signal Process..

[20]  H. Zou.  The Adaptive Lasso and Its Oracle Properties , 2006 .

[21]  Francis R. Bach,et al.  Bolasso: model consistent Lasso estimation through the bootstrap , 2008, ICML '08.

[22]  Chih-Jen Lin,et al.  LIBSVM: A library for support vector machines , 2011, TIST.

[23]  M. Yuan,et al.  Model selection and estimation in regression with grouped variables , 2006 .

[24]  Jill P. Mesirov,et al.  Consensus Clustering: A Resampling-Based Method for Class Discovery and Visualization of Gene Expression Microarray Data , 2003, Machine Learning.

[25]  Jean-Philippe Vert,et al.  The Influence of Feature Selection Methods on Accuracy, Stability and Interpretability of Molecular Signatures , 2011, PloS one.

[26]  Joel A. Tropp,et al.  Greed is good: algorithmic results for sparse approximation , 2004, IEEE Transactions on Information Theory.

[27]  Russell A. Poldrack,et al.  Analyses of regional-average activation and multivoxel pattern information tell complementary stories , 2012, Neuropsychologia.

[28]  Peng Zhao,et al.  On Model Selection Consistency of Lasso , 2006, J. Mach. Learn. Res..

[29]  Francis R. Bach,et al.  Trace Lasso: a trace norm regularization for correlated designs , 2011, NIPS.