This package combines different tools created for the alalysis of the Cambridge Center for Aging and Neuroscience (Cam-CAN) data set. Structural MRI data should be preprocessed using FreeSurfer. To prepare fMRI files use pypreprocess, a sample configuration can be found in camcan_analysis/scripts/preprocessing
- FreeSurfer
- scipy>=0.9
- numpy>=1.6.1
- scikit-learn>=0.14.1
- joblib>=0.11.0
- nibabel>=1.2.0
- nilearn>=0.3.0
- nistats
- pandas>=0.19.2
The camcan directory holds tools for data analysis.
The data folder contains information genrearated by processing scripts and some of the dat provided with the Cam-CAN data set. Information about subjects' age can be found in data/participant_data.csv.
Jupyter notebooks with example of the brain age prediction can be found in camcan_analysis/notebook/age_prediction/
To analyze the data use files from the scripts folder.
scripts/feature_extraction contains files for feature extraction from structural and functional MRI, MEG data:
- combine_connectivity.py - extract connectivity information from timeseries and save it into a single file
- extract_connectivity.py - extract and save connectivity from fMRI timeseries
- extract_timeseries.py - extract and save timeseries from fMRI volumes
- extract_truncated_connectivity.py - extract and save connectivity from fMRI timeseries of different duration
- extract_structural.py - extract and save structural data from FreeSurfer outputs
scripts/plotting contains utility code for producing various plots using maplotlib, bqplot, etc. To see examples check data/figures.
To train and test age prediction models use files from the scripts/prediction directory:
- age_prediction_meg_source_space.py - age prediction using MRI, fMRI and MEG data (source space, etc.)
- age_prediciton_mri_meg_spoc.py - age prediction using MRI, fMRI and MEG data. MEG data is processed using the SPOC algorithm, see implementation details in camcan_analysis/camcan/processing/spoc.py
- age_prediction_trunc_fmri.py - age prediction using fMRI timeseries of different length