The term Searchlight refers to an algorithm that runs a scalar Measure on all possible spheres of a certain size within a dataset (that provides information about distances between feature locations). The measure typically computed is a cross-validation of a classifier performance (see CrossValidation section in the tutorial). The idea to use a searchlight as a sensitivity analyzer on fMRI datasets stems from Kriegeskorte et al. (2006).
A searchlight analysis is can be easily performed. This examples shows a minimal draft of a complete analysis.
First import a necessary pieces of PyMVPA – this time each bit individually.
import numpy as np
from mvpa2.generators.partition import OddEvenPartitioner
from mvpa2.clfs.svm import LinearCSVMC
from mvpa2.measures.base import CrossValidation
from mvpa2.measures.searchlight import sphere_searchlight
from mvpa2.testing.datasets import datasets
from mvpa2.mappers.fx import mean_sample
For the sake of simplicity, let’s use a small artificial dataset.
# Lets just use our tiny 4D dataset from testing battery
dataset = datasets['3dlarge']
Now it only takes three lines for a searchlight analysis.
# setup measure to be computed in each sphere (cross-validated
# generalization error on odd/even splits)
cv = CrossValidation(LinearCSVMC(), OddEvenPartitioner())
# setup searchlight with 2 voxels radius and measure configured above
sl = sphere_searchlight(cv, radius=2, space='myspace',
postproc=mean_sample())
# run searchlight on dataset
sl_map = sl(dataset)
print 'Best performing sphere error:', np.min(sl_map.samples)
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_minimal.py).
