/VGDL-fMRI-Python-Data-Analysis

using brainIAK

Primary LanguageJupyter Notebook

fMRI Data Analysis in Python

This repository contains Python code and Jupyter notebooks used to analyse fMRI data.

What's in the folder: Analyses_Daphne? 🤔

This folder contains the deliverables from Daphne, as result of the summer research assistantship with Momchil from Jun-Aug 2020. You will find code on preprocessing data (🟠), how to find regions of interest (🔴), intersubject correlation (🟢) and functional connectivity analyses (🔵).

If anything is unclear, feel free to send me an email.

Python scripts

  • utils.py includes all the functions used for preprocessing the data, performing analyses, plotting and more. All the notebooks in this folder use the functions from utils.py to keep them nice and clean.

Notebooks

(almost) each of the notebooks include a summary of what is done and are well documented. Some also include links that point to resources where you can learn more about the respective analysis.

Notebooks where I preprocess data or just stored for reference

  • Preprocess_smooth_betas.ipynb is where the .mat files are preprocessed and the BOLD data is reordered. We looked at beta series for boxcars, blocks, levels, and games. 🟠

  • Startingpoint.ipynb is probably not very useful but kept it here for potential debugging.

Analysis notebooks

  • ISC_nosmooth.ipynb Doing an intersubject correlation (ISC) with the nonsmooth betas (row wise). 🟢

  • My_ISC_vs_brainiak_ISC.ipynb Here we compare the results obtained from our own implemented intersubject correlation with the results from brainiaks intersubject correlation (from brainiak.isc import isc) 🟢

  • Smooth_betas_statisticalMaps.ipynb This is were the statistical maps for the slides are generated. We perform one-sample t-tests and use the functions from utils.py to plot the t statistics on the anatomical brain image. 🟢

  • In From_vox_to_corrs.ipynb we look at how the intersubject correlations evolve across levels. 🟢

  • Find_ROIs.ipynb This notebook is used to find the regions of interest (ROIs) based on the statistical maps and extract the most intense voxel for each of the found ROIs. 🔴

  • In Running_on_cluster.ipynb we get the data for all the subjects later used in Functional_connectivity_within_subjects.ipynb. 🔵

  • Functional_connectivity_within_subjects.ipynb does a standard FC analysis, correlating across different brain areas (i.e. the ROIs found by Momchil & the ones from me) within a subjects brain. It also includes a bit on how to perform granger causality on voxel pairs. 🔵

  • Functional_connectivity_between_subjects.ipynb as the name suggests, here we do the look at the functional connectivity across subjects. 🔵