fmri.smooth {fmri}R Documentation

Smoothing Statistical Parametric Maps

Description

Perform the adaptive weights smoothing procedure

Usage

fmri.smooth(spm, hmax = 4, adaptive=TRUE,
            lkern="Triangle", skern="Triangle", na.rm=FALSE)

Arguments

spm object of class fmrispm
hmax maximum bandwidth to smooth
adaptive logical. TRUE (default) for adaptive smoothing
lkern lkern specifies the location kernel. Defaults to "Triangle", other choices are "Gaussian", "Quadratic", "Cubic" and "Uniform". Note that the location kernel is applied to (x-x_j)^2/h^2, i.e. the use of "Triangle" corresponds to the Epanechnicov kernel in nonparametric kernel regression.
skern skern specifies the kernel for the statistical penalty. Defaults to "Triangle", the alternative is "Exp". lkern="Triangle" allows for much faster computation (saves up to 50%).
na.rm na.rm specifies how NA's in the SPM are handled. NA's may occur in voxel where the time series information did not allow for estimating parameters and their variances or where the time series information where constant over time. A high (1e19) value of the variance and a parameter of 0 are used to characterize NA's. If na.rm=FALSE these values are simply downweigthed by their high variance and estimates in these voxels are produced as nonadaptive averages of values from neighboring voxels. If na.rm=TRUE no estimates are computed (the values for mean and variance are kept) in these voxels.

Details

This function performs the smoothing on the Statistical Parametric Map spm.

hmax is the (maximal) bandwidth used in the last iteration. Choose adaptive as FALSE for non adaptive smoothing. lkern can be used for specifying the localization kernel. For comparison with non adaptive methods use "Gaussian" (hmax given in FWHM), for better adaptation use "Triangle" (default, hmax given in voxel). skern can be used for specifying the kernel for the statistical penalty.

The function handles zero variances by assigning a large value (1e20) to these variances.

Value

object with class attributes "fmrispm" and "fmridata"

cbeta smoothed parameter estimate
var variance of the parameter
hmax maximum bandwidth used
rxyz smoothness in resel space. all directions
rxyz0 smoothness in resel space as would be achieved by a Gaussian filter with the same bandwidth. all directions
scorr spatial correlation of original data
weights ratio of voxel dimensions
vwghts ratio of estimated variances for the stimuli given by vvector
hrf Expected BOLD response for the specified effect

Author(s)

Joerg Polzehl polzehl@wias-berlin.de, Karsten Tabelow tabelow@wias-berlin.de

References

Tabelow, K., Polzehl, J., Voss, H.U., and Spokoiny, V. (2005). Analysing {fMRI} experiments with structure adaptive smoothing procedures, NeuroImage, accepted (2006).

Polzehl, J. and Spokoiny, V. (2006). Propagation-Separation Approach for Local Likelihood Estimation, Probab. Theory Relat. Fields 135, 335-362.

Examples

## Not run: fmri.smooth(spm, hmax = 4, lkern = "Gaussian")

[Package fmri version 1.01 Index]