NeuroConn is a Python package that provides a user-friendly interface for fMRI preprocessing and computing connectivity matrices and gradients. It is designed as a BIDS application, allowing easy integration with BIDS-formatted datasets. Documentation: https://victoris93.github.io/NeuroConn/
NB! If you wish to run fmriprep within this package, install Docker Desktop first. Keep it running when you start RawDataset.docker_fmriprep()
- Preprocessing of fMRI data using the fmriprep pipeline
- Computation of connectivity matrices and gradients
- Direct output of gradients or connectivity matrices for any subject without specifying preprocessing parameters
- Handling of BIDS-formatted datasets
You can install NeuroConn using pip: pip install NeuroConn
1. fMRIPrep. The class RawDataset features a method to run fmriprep within within your Python environment. Before running it:
- Register with freesurfer and download the license file
freesurfer_license.txt) - Install Docker Desktop.
- After having activated your environment, run
pip install fmriprep-docker. - Start Docker Desktop. Then, give this a try:
from NeuroConn.preprocessing.preprocessing import RawDataset, FmriPreppedDataSet
from NeuroConn.data.example_datasets import fetch_example_data
ex_data = fetch_example_data() # from https://openneuro.org/datasets/ds002748
data = RawDataset(ex_data)
subject = '52'
data.docker_fmriprep(subject, fs_reconall = False, fs_license = <path_to_freesurfer_license.txt>)
2. Post-fMRIPrep Here's an example of how to use the FmriPreppedDataSet class provided by NeuroConn:
from NeuroConn.preprocessing.preprocessing import RawDataset, FmriPreppedDataSet
from NeuroConn.data.example_datasets import fetch_example_data
# Download the dataset preprocessed with fMRIPrep
example_data = fetch_example_data('https://drive.google.com/file/d/1XjF5wDJXHzMyfoAjQE6NW2xcj9PulZzH/view?usp=share_link')
# Initialize the dataset object
dataset = FmriPreppedDataSet(example_data)
# Compute connectivity matrix
conn_matrix = data_prepped.get_conn_matrix(subject, parcellation='schaefer', task='rest', n_parcels=1000, save = True)
# Compute 10 gradients (Margulies et al., 2016)
gradients = get_gradients(data_prepped,subject, task='rest', n_components = 10, approach = "pca")
For more detailed information and examples, please refer to the notebook.
Contributions are welcome! If you find any issues or have suggestions for improvements, please open an issue or submit a pull request on this GitHub repository.
NeuroConn is released under the MIT License. See the LICENSE file for more details.
Bezmaternykh D.D., Melnikov M.Y., Savelov A.A. et al. Brain Networks Connectivity in Mild to Moderate Depression: Resting State fMRI Study with Implications to Nonpharmacological Treatment. Neural Plasticity, 2021. V. 2021. № 8846097. PP. 1-15. DOI: 10.1155/2021/8846097