fMRIDenoise - automated denoising, denoising strategies comparison, and functional connectivity data quality control.

Tool for automatic denoising, denoising strategies comparisons, and functional connectivity data quality control. The goal of fMRIDenoise is to provide an objective way to select best-performing denoising strategy given the data. FMRIDenoise is designed to work directly on fMRIPrep-preprocessed datasets and data in BIDS standard. We believe that the tool can make the selection of the denoising strategy more objective and also help researchers to obtain FC quality control metrics with almost no effort.

The project is in alpha stage and we are looking for feedback and collaborators.

Problem

Solution

Installation

In a project directory run:

python setup.py install (--user)

To install fmridenoise from PyPi run:

pip install fmridenoise (--user)

Execution

fmridenoise or python -m fmridenoise

usage: fmridenoise [-h] [-sub SUBJECTS [SUBJECTS ...]]
                [-ses SESSIONS [SESSIONS ...]] [-t TASKS [TASKS ...]]
                [-p PIPELINES [PIPELINES ...]]
                [-d DERIVATIVES [DERIVATIVES ...]] [--high-pass HIGH_PASS]
                [--low-pass LOW_PASS] [--use-aroma] [--MultiProc] [--profiler PROFILER]
                [-g] [--graph GRAPH] [--dry]
                bids_dir

positional arguments:
bids_dir                Path do preprocessed BIDS dataset.

optional arguments:
-h, --help              Show help message and exit.
-sub SUBJECTS [SUBJECTS ...], --subjects SUBJECTS [SUBJECTS ...]
                        List of subjects
-ses SESSIONS [SESSIONS ...], --sessions SESSIONS [SESSIONS ...]
                        List of session numbers, separated with spaces.
-t TASKS [TASKS ...], --tasks TASKS [TASKS ...]
                        List of tasks names, separated with spaces.
-p PIPELINES [PIPELINES ...], --pipelines PIPELINES [PIPELINES ...]
                        Name of pipelines used for denoising, can be both
                        paths to json files with pipeline or name of pipelines
                        from package.
-d DERIVATIVES [DERIVATIVES ...], --derivatives DERIVATIVES [DERIVATIVES ...]
                        Name (or list) of derivatives for which fmridenoise
                        should be run. By default workflow looks for fmriprep
                        dataset.
--high-pass HIGH_PASS
                        High pass filter value, deafult 0.008.
--low-pass LOW_PASS     Low pass filter value, default 0.08
--use-aroma             Run denoising pipelines based on ICA-AROMA, default False
--MultiProc             Run script on multiple processors, default False
--profiler PROFILER     Run profiler along workflow execution to estimate
                        resources usage PROFILER is path to output log file.
-g, --debug             Run fmridenoise in debug mode - richer output, stops
                        on first unchandled exception.
--graph GRAPH           Create workflow graph at GRAPH path
--dry                   Perform everything except actually running workflow