Brain kernel: A new spatial covariance function for fMRI data

A key problem in functional magnetic resonance imaging (fMRI) is to estimate spatial activity patterns from noisy high-dimensional signals. Spatial smoothing provides one approach to regularizing such estimates. However, standard smoothing methods ignore the fact that correlations in neural activity...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Anqi Wu, Samuel A. Nastase, Christopher A. Baldassano, Nicholas B. Turk-Browne, Kenneth A. Norman, Barbara E. Engelhardt, Jonathan W. Pillow
Formato:	article
Lenguaje:	EN
Publicado:	Elsevier 2021
Materias:	Brain kernel Gaussian process Latent variable model Brain decoding Factor modeling Resting-state fmri Neurosciences. Biological psychiatry. Neuropsychiatry RC321-571
Acceso en línea:	https://doaj.org/article/7953409cdc8143cca57ed3d4ffd1a406
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:7953409cdc8143cca57ed3d4ffd1a406
record_format	dspace
spelling	oai:doaj.org-article:7953409cdc8143cca57ed3d4ffd1a4062021-11-10T04:21:08ZBrain kernel: A new spatial covariance function for fMRI data1095-957210.1016/j.neuroimage.2021.118580https://doaj.org/article/7953409cdc8143cca57ed3d4ffd1a4062021-12-01T00:00:00Zhttp://www.sciencedirect.com/science/article/pii/S1053811921008533https://doaj.org/toc/1095-9572A key problem in functional magnetic resonance imaging (fMRI) is to estimate spatial activity patterns from noisy high-dimensional signals. Spatial smoothing provides one approach to regularizing such estimates. However, standard smoothing methods ignore the fact that correlations in neural activity may fall off at different rates in different brain areas, or exhibit discontinuities across anatomical or functional boundaries. Moreover, such methods do not exploit the fact that widely separated brain regions may exhibit strong correlations due to bilateral symmetry or the network organization of brain regions. To capture this non-stationary spatial correlation structure, we introduce the brain kernel, a continuous covariance function for whole-brain activity patterns. We define the brain kernel in terms of a continuous nonlinear mapping from 3D brain coordinates to a latent embedding space, parametrized with a Gaussian process (GP). The brain kernel specifies the prior covariance between voxels as a function of the distance between their locations in embedding space. The GP mapping warps the brain nonlinearly so that highly correlated voxels are close together in latent space, and uncorrelated voxels are far apart. We estimate the brain kernel using resting-state fMRI data, and we develop an exact, scalable inference method based on block coordinate descent to overcome the challenges of high dimensionality (10-100K voxels). Finally, we illustrate the brain kernel’s usefulness with applications to brain decoding and factor analysis with multiple task-based fMRI datasets.Anqi WuSamuel A. NastaseChristopher A. BaldassanoNicholas B. Turk-BrowneKenneth A. NormanBarbara E. EngelhardtJonathan W. PillowElsevierarticleBrain kernelGaussian processLatent variable modelBrain decodingFactor modelingResting-state fmriNeurosciences. Biological psychiatry. NeuropsychiatryRC321-571ENNeuroImage, Vol 245, Iss , Pp 118580- (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Brain kernel Gaussian process Latent variable model Brain decoding Factor modeling Resting-state fmri Neurosciences. Biological psychiatry. Neuropsychiatry RC321-571
spellingShingle	Brain kernel Gaussian process Latent variable model Brain decoding Factor modeling Resting-state fmri Neurosciences. Biological psychiatry. Neuropsychiatry RC321-571 Anqi Wu Samuel A. Nastase Christopher A. Baldassano Nicholas B. Turk-Browne Kenneth A. Norman Barbara E. Engelhardt Jonathan W. Pillow Brain kernel: A new spatial covariance function for fMRI data
description	A key problem in functional magnetic resonance imaging (fMRI) is to estimate spatial activity patterns from noisy high-dimensional signals. Spatial smoothing provides one approach to regularizing such estimates. However, standard smoothing methods ignore the fact that correlations in neural activity may fall off at different rates in different brain areas, or exhibit discontinuities across anatomical or functional boundaries. Moreover, such methods do not exploit the fact that widely separated brain regions may exhibit strong correlations due to bilateral symmetry or the network organization of brain regions. To capture this non-stationary spatial correlation structure, we introduce the brain kernel, a continuous covariance function for whole-brain activity patterns. We define the brain kernel in terms of a continuous nonlinear mapping from 3D brain coordinates to a latent embedding space, parametrized with a Gaussian process (GP). The brain kernel specifies the prior covariance between voxels as a function of the distance between their locations in embedding space. The GP mapping warps the brain nonlinearly so that highly correlated voxels are close together in latent space, and uncorrelated voxels are far apart. We estimate the brain kernel using resting-state fMRI data, and we develop an exact, scalable inference method based on block coordinate descent to overcome the challenges of high dimensionality (10-100K voxels). Finally, we illustrate the brain kernel’s usefulness with applications to brain decoding and factor analysis with multiple task-based fMRI datasets.
format	article
author	Anqi Wu Samuel A. Nastase Christopher A. Baldassano Nicholas B. Turk-Browne Kenneth A. Norman Barbara E. Engelhardt Jonathan W. Pillow
author_facet	Anqi Wu Samuel A. Nastase Christopher A. Baldassano Nicholas B. Turk-Browne Kenneth A. Norman Barbara E. Engelhardt Jonathan W. Pillow
author_sort	Anqi Wu
title	Brain kernel: A new spatial covariance function for fMRI data
title_short	Brain kernel: A new spatial covariance function for fMRI data
title_full	Brain kernel: A new spatial covariance function for fMRI data
title_fullStr	Brain kernel: A new spatial covariance function for fMRI data
title_full_unstemmed	Brain kernel: A new spatial covariance function for fMRI data
title_sort	brain kernel: a new spatial covariance function for fmri data
publisher	Elsevier
publishDate	2021
url	https://doaj.org/article/7953409cdc8143cca57ed3d4ffd1a406
work_keys_str_mv	AT anqiwu brainkernelanewspatialcovariancefunctionforfmridata AT samuelanastase brainkernelanewspatialcovariancefunctionforfmridata AT christopherabaldassano brainkernelanewspatialcovariancefunctionforfmridata AT nicholasbturkbrowne brainkernelanewspatialcovariancefunctionforfmridata AT kennethanorman brainkernelanewspatialcovariancefunctionforfmridata AT barbaraeengelhardt brainkernelanewspatialcovariancefunctionforfmridata AT jonathanwpillow brainkernelanewspatialcovariancefunctionforfmridata
_version_	1718440663943479296

Brain kernel: A new spatial covariance function for fMRI data

Ejemplares similares