mvpa2.algorithms.searchlight_hyperalignment.compute_feature_scores(datasets, exclude_from_model=None)

Takes a list of datasets and computes a magical feature score for each feature in each dataset Haxby et al., Neuron (2011)


datasets : list or tuple of datasets

exclude_from_model: list of dataset indices that won’t participate

in voxel selection of others


list : a list of feature scores; higher the score, better the feature

NOTE: This function assumes that the datasets are zscored