fMRIPost-tedana

A BIDS App for the tedana workflow

Overview

fMRIPost-tedana post-processes multi-echo fMRI data that have been minimally preprocessed with fMRIPrep. In addition to the basic multi-echo independent components analysis (ME-ICA) performed by tedana, fMRIPost-tedana will leverage fMRIPrep derivatives, like extracted confounds and high-resolution tissue masks, to improve the multi-echo denoising performance.

Workflow

1. A BIDS data grabber to collect the first echo from each run.
2. A BIDS data grabber to grab the other echoes, brain mask, echo times, and potentially confounds.
3. Drop dummy volumes as requested.
4. Run tedana.
   • Limit to the brain mask.
   • Extract echo times from metadata.
   • Include fMRIPrep confounds as external regressors.
   • Incorporate tissue type masks in decision tree?
5. Supplement ME-ICA with other component classification algorithms.
   • AROMA?
   • CTLP?
6. Reorganize tedana outputs into BIDS format.
7. Denoise optimally combined data in target spaces.
8. Generate figures summarizing the denoising results.
   • Carpet plots of denoised data
   • T2* map
   • S0 map
   • Component weight maps, time series, periodograms, and classifications
   • Components-by-confound correlation heat map
9. Generate nireports HTML report.