Structured sparsity penalization has recently improved statistical models applied to high-dimensional data in various domains. As an extension to medical imaging, the present work incorporates priors on network hierarchies of brain regions into logistic-regression to distinguish neural activity effects. These priors bridge two separately studied levels of brain architecture: functional segregation into regions and functional integration by networks. Hierarchical region-network priors are shown to better classify and recover 18 psychological tasks than other sparse estimators. Varying the relative importance of region and network structure within the hierarchical tree penalty captured complementary aspects of the neural activity patterns. Local and global priors of neurobiological knowledge are thus demonstrated to offer advantages in generalization performance, sample complexity, and domain interpretability.
%0 Book Section
%1 bzdok2017hierarchical
%A Bzdok, Danilo
%A Eickenberg, Michael
%A Varoquaux, Gaël
%A Thirion, Bertrand
%B Information Processing in Medical Imaging
%D 2017
%E Niethammer, Marc
%E Styner, Martin
%E Aylward, Stephen
%E Zhu, Hongtu
%E Oguz, Ipek
%E Yap, Pew-Thian
%E Shen, Dinggang
%I Springer International Publishing
%K machine_learning neuroscience structured_sparsity
%P 323--335
%R 10.1007/978-3-319-59050-9_26
%T Hierarchical Region-Network Sparsity for High-Dimensional Inference in Brain Imaging.
%U http://dx.doi.org/10.1007/978-3-319-59050-9_26
%V 10265
%X Structured sparsity penalization has recently improved statistical models applied to high-dimensional data in various domains. As an extension to medical imaging, the present work incorporates priors on network hierarchies of brain regions into logistic-regression to distinguish neural activity effects. These priors bridge two separately studied levels of brain architecture: functional segregation into regions and functional integration by networks. Hierarchical region-network priors are shown to better classify and recover 18 psychological tasks than other sparse estimators. Varying the relative importance of region and network structure within the hierarchical tree penalty captured complementary aspects of the neural activity patterns. Local and global priors of neurobiological knowledge are thus demonstrated to offer advantages in generalization performance, sample complexity, and domain interpretability.
@incollection{bzdok2017hierarchical,
abstract = {Structured sparsity penalization has recently improved statistical models applied to high-dimensional data in various domains. As an extension to medical imaging, the present work incorporates priors on network hierarchies of brain regions into logistic-regression to distinguish neural activity effects. These priors bridge two separately studied levels of brain architecture: functional segregation into regions and functional integration by networks. Hierarchical region-network priors are shown to better classify and recover 18 psychological tasks than other sparse estimators. Varying the relative importance of region and network structure within the hierarchical tree penalty captured complementary aspects of the neural activity patterns. Local and global priors of neurobiological knowledge are thus demonstrated to offer advantages in generalization performance, sample complexity, and domain interpretability.},
added-at = {2018-12-07T09:10:16.000+0100},
author = {Bzdok, Danilo and Eickenberg, Michael and Varoquaux, Gaël and Thirion, Bertrand},
biburl = {https://www.bibsonomy.org/bibtex/230c326e0181a2f5b340e99fbf71ee3c8/jpvaldes},
booktitle = {Information Processing in Medical Imaging},
citeulike-article-id = {14571732},
citeulike-linkout-0 = {http://dx.doi.org/10.1007/978-3-319-59050-9\_26},
citeulike-linkout-1 = {http://link.springer.com/chapter/10.1007/978-3-319-59050-9\_26},
doi = {10.1007/978-3-319-59050-9_26},
editor = {Niethammer, Marc and Styner, Martin and Aylward, Stephen and Zhu, Hongtu and Oguz, Ipek and Yap, Pew-Thian and Shen, Dinggang},
interhash = {7743ec894991c81d1fc8631b0ccd80f0},
intrahash = {30c326e0181a2f5b340e99fbf71ee3c8},
keywords = {machine_learning neuroscience structured_sparsity},
pages = {323--335},
posted-at = {2018-04-19 13:05:39},
priority = {2},
publisher = {Springer International Publishing},
series = {Lecture Notes in Computer Science},
timestamp = {2018-12-07T11:27:26.000+0100},
title = {Hierarchical Region-Network Sparsity for High-Dimensional Inference in Brain Imaging.},
type = {Publication},
url = {http://dx.doi.org/10.1007/978-3-319-59050-9_26},
volume = 10265,
year = 2017
}