Limits...
Topographic factor analysis: a Bayesian model for inferring brain networks from neural data.

Manning JR, Ranganath R, Norman KA, Blei DM - PLoS ONE (2014)

Bottom Line: The neural patterns recorded during a neuroscientific experiment reflect complex interactions between many brain regions, each comprising millions of neurons.However, the measurements themselves are typically abstracted from that underlying structure.Specifically, TFA casts each brain image as a weighted sum of spatial functions.

View Article: PubMed Central - PubMed

Affiliation: Princeton Neuroscience Institute, Princeton University, Princeton, New Jersey, United States of America; Department of Computer Science, Princeton University, Princeton, New Jersey, United States of America.

ABSTRACT
The neural patterns recorded during a neuroscientific experiment reflect complex interactions between many brain regions, each comprising millions of neurons. However, the measurements themselves are typically abstracted from that underlying structure. For example, functional magnetic resonance imaging (fMRI) datasets comprise a time series of three-dimensional images, where each voxel in an image (roughly) reflects the activity of the brain structure(s)-located at the corresponding point in space-at the time the image was collected. FMRI data often exhibit strong spatial correlations, whereby nearby voxels behave similarly over time as the underlying brain structure modulates its activity. Here we develop topographic factor analysis (TFA), a technique that exploits spatial correlations in fMRI data to recover the underlying structure that the images reflect. Specifically, TFA casts each brain image as a weighted sum of spatial functions. The parameters of those spatial functions, which may be learned by applying TFA to an fMRI dataset, reveal the locations and sizes of the brain structures activated while the data were collected, as well as the interactions between those structures.

Show MeSH
Decomposing a brain image into a weighted combination of sources.A coronal slice from an example brain image is shown on the left. TFA approximates the image as a weighted sum of source images. The approximation (reconstruction) is shown in the middle panel, and several of the (weighted) source images are shown on the right. The color scale is the same as for Figure 1.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4012983&req=5

pone-0094914-g002: Decomposing a brain image into a weighted combination of sources.A coronal slice from an example brain image is shown on the left. TFA approximates the image as a weighted sum of source images. The approximation (reconstruction) is shown in the middle panel, and several of the (weighted) source images are shown on the right. The color scale is the same as for Figure 1.

Mentions: TFA assumes that fMRI images reflect the activities of a finite number of sources distributed throughout the brain (Figure 2). (This is a simplifying assumption, of course, but is useful for uncovering hidden structures in brain activity data.) Intuitively, a source could reflect a particular brain structure, or a set of nearby brain structures behaving similarly or carrying out similar computations during an experiment. Each source in TFA is formally defined by a set of parameters of a spatial function. In principle, we may choose any family of spatial functions that describes the sources’ shapes (see Discussion). To simplify the presentation, sources in our implementation will be specified as sets of parameters of Gaussian radial basis functions (RBFs). If an RBF has center and (log) width , then its activation at location is given by:(1)


Topographic factor analysis: a Bayesian model for inferring brain networks from neural data.

Manning JR, Ranganath R, Norman KA, Blei DM - PLoS ONE (2014)

Decomposing a brain image into a weighted combination of sources.A coronal slice from an example brain image is shown on the left. TFA approximates the image as a weighted sum of source images. The approximation (reconstruction) is shown in the middle panel, and several of the (weighted) source images are shown on the right. The color scale is the same as for Figure 1.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4012983&req=5

pone-0094914-g002: Decomposing a brain image into a weighted combination of sources.A coronal slice from an example brain image is shown on the left. TFA approximates the image as a weighted sum of source images. The approximation (reconstruction) is shown in the middle panel, and several of the (weighted) source images are shown on the right. The color scale is the same as for Figure 1.
Mentions: TFA assumes that fMRI images reflect the activities of a finite number of sources distributed throughout the brain (Figure 2). (This is a simplifying assumption, of course, but is useful for uncovering hidden structures in brain activity data.) Intuitively, a source could reflect a particular brain structure, or a set of nearby brain structures behaving similarly or carrying out similar computations during an experiment. Each source in TFA is formally defined by a set of parameters of a spatial function. In principle, we may choose any family of spatial functions that describes the sources’ shapes (see Discussion). To simplify the presentation, sources in our implementation will be specified as sets of parameters of Gaussian radial basis functions (RBFs). If an RBF has center and (log) width , then its activation at location is given by:(1)

Bottom Line: The neural patterns recorded during a neuroscientific experiment reflect complex interactions between many brain regions, each comprising millions of neurons.However, the measurements themselves are typically abstracted from that underlying structure.Specifically, TFA casts each brain image as a weighted sum of spatial functions.

View Article: PubMed Central - PubMed

Affiliation: Princeton Neuroscience Institute, Princeton University, Princeton, New Jersey, United States of America; Department of Computer Science, Princeton University, Princeton, New Jersey, United States of America.

ABSTRACT
The neural patterns recorded during a neuroscientific experiment reflect complex interactions between many brain regions, each comprising millions of neurons. However, the measurements themselves are typically abstracted from that underlying structure. For example, functional magnetic resonance imaging (fMRI) datasets comprise a time series of three-dimensional images, where each voxel in an image (roughly) reflects the activity of the brain structure(s)-located at the corresponding point in space-at the time the image was collected. FMRI data often exhibit strong spatial correlations, whereby nearby voxels behave similarly over time as the underlying brain structure modulates its activity. Here we develop topographic factor analysis (TFA), a technique that exploits spatial correlations in fMRI data to recover the underlying structure that the images reflect. Specifically, TFA casts each brain image as a weighted sum of spatial functions. The parameters of those spatial functions, which may be learned by applying TFA to an fMRI dataset, reveal the locations and sizes of the brain structures activated while the data were collected, as well as the interactions between those structures.

Show MeSH