tedana.selection.tedica.automatic_selection

automatic_selection(component_table, n_echos, n_vols, tree='kundu', verbose=False)

Classify components based on component table and decision tree type.

Parameters:

• component_table (pd.DataFrame) – The component table to classify

• n_echos (int) – The number of echoes in this dataset

• tree (str) – The type of tree to use for the ComponentSelector object. Default=”kundu”

• verbose (bool) – More verbose logging output if True. Default=False

Returns:

selector (tedana.selection.component_selector.ComponentSelector) – Contains component classifications in a component_table and provenance and metadata from the component selection process

Notes

If tree=kundu, the selection algorithm used in this function was originated in ME-ICA by Prantik Kundu, and his original implementation is available at: https://github.com/ME-ICA/me-ica/blob/ b2781dd087ab9de99a2ec3925f04f02ce84f0adc/meica.libs/select_model.py

The appropriate citation is Kundu et al.[1].

This component selection process uses multiple, previously calculated metrics that include kappa, rho, variance explained, noise and spatial frequency metrics, and measures of spatial overlap across metrics.

Prantik began to update these selection criteria to use SVMs to distinguish components, a hypercommented version of this attempt is available at: https://gist.github.com/emdupre/ca92d52d345d08ee85e104093b81482e

If tree==”minimal”, a selection algorithm based on the “kundu” tree will be used. The differences between the “minimal” and “kundu” trees are described in the FAQ.

References

[1]

Prantik Kundu, Noah D Brenowitz, Valerie Voon, Yulia Worbe, Petra E Vértes, Souheil J Inati, Ziad S Saad, Peter A Bandettini, and Edward T Bullmore. Integrated strategy for improving functional connectivity mapping using multiecho fmri. Proceedings of the National Academy of Sciences, 110(40):16187–16192, 2013. URL: https://doi.org/10.1073/pnas.1301725110, doi:10.1073/pnas.1301725110.