This content refers to the previous stable release of PyMVPA. Please visit www.pymvpa.org for the most recent version of PyMVPA and its documentation.

A searchlight computing a dissimilarity matrix measure

This example extends the minimal Searchlight example to use a dissimilarity matrix-based DatasetMetric to compute Searchlight-center significance. This is based on representational similarity analysis (RSA) as presented in Kriegeskorte et al. (2008).

First import all necessary parts of PyMVPA.

from mvpa.suite import *

Create a small artificial dataset.

# overcomplicated way to generate an example dataset
ds = normalFeatureDataset(perlabel=10, nlabels=2, nchunks=2,
                          nfeatures=10, nonbogus_features=[3, 7],
                          snr=5.0)
dataset = MaskedDataset(samples=ds.samples, labels=ds.labels,
                        chunks=ds.chunks)

Create a dissimilarity matrix based on the labels of the data points in our test dataset. This will allow us to see if there is a correlation between any given searchlight sphere and the experimental conditions.

# create dissimilarity matrix using the 'confusion' distance
# metric
dsm = DSMatrix(dataset.labels, 'confusion')

Now it only takes three lines for a searchlight analysis.

# setup measure to be computed in each sphere (correlation
# distance between dissimilarity matrix and the dissimilarities
# of a particular searchlight sphere across experimental
# conditions), N.B. in this example between-condition
# dissimilarity is also pearson's r (i.e., correlation distance)
dsmetric = DSMDatasetMeasure(dsm, 'pearson', 'pearson')

# setup searchlight with 5 mm radius and measure configured above
sl = Searchlight(dsmetric, radius=5)

# run searchlight on dataset
sl_map = sl(dataset)

print 'Best performing sphere error:', max(sl_map)

If this analysis is done on a fMRI dataset using NiftiDataset the resulting searchlight map (sl_map) can be mapped back into the original dataspace and viewed as a brain overlay. Another example shows a typical application of this algorithm.

See also

The full source code of this example is included in the PyMVPA source distribution (doc/examples/searchlight_dsm.py).