N2G.Spatial.Mixture {AnalyzeFMRI} R Documentation

## fMRI Spatial Mixture Modelling

### Description

Fits the spatial mixture model of Hartvig and Jensen (2000)

### Usage

```N2G.Spatial.Mixture(data, par.start = c(4, 2, 4, 2, 0.9, 0.05),
ksize, ktype = c("2D", "3D"), mask = NULL)
```

### Arguments

 `data` The dataset (usually a vector) `par.start` Starting values for N2G model `ksize` Kernel size (see paper) `ktype` Format of kernel "2D" or "3D" `mask` Mask for dataset.

### Value

p.map = a1, par = fit\$par, lims = fit\$lims Returns a list with following components

 `p.map` Posterior Probability Map of activation `par` Fitted parameters of the underlying N2G model `lims` Normal component interval for fitted model

J. L. Marchini

### References

Hartvig and Jensen (2000) Spatial Mixture Modelling of fMRI Data

`N2G.Class.Probability`, `N2G.Likelihood.Ratio`, `N2G.Density` , `N2G.Likelihood` , `N2G.Transform`, `N2G.Fit` , `N2G` , `N2G.Inverse` , `N2G.Region`

### Examples

```
## simulate image
d <- c(100, 100, 1)
y <- array(0, dim = d)
m <- y
m[, , ] <- 1

z.init <- 2 * m
z.init[20:40, 20:40, 1] <- 1
z.init[50:70, 50:70, 1] <- 3

y[z.init == 1] <- -rgamma(sum(z.init == 1), 4, 1)
y[z.init == 2] <- rnorm(sum(z.init == 2))
y[z.init == 3] <- rgamma(sum(z.init == 3), 4, 1)

mask <- 1 * (y < 1000)

## fit spatial mixture model
ans <- N2G.Spatial.Mixture(y, par.start = c(4, 2, 4, 2, 0.9, 0.05),
ksize = 3, ktype = "2D", mask = m)

## plot original image, standard mixture model estimate and spatial mixture
## model estimate

par(mfrow = c(1, 3))
image(y[, , 1])
image(y[, , 1] > ans\$lims[1]) ## this line plots the results of a Non-Spatial Mixture Model
image(ans\$p.map[, , 1] > 0.5) ## this line plots the results of the Spatial Mixture Model

```

[Package AnalyzeFMRI version 1.1-24 Index]