Stroke-EEG-Brain-network-analysis

Functional connectivity and brain network (graph theory) analysis for motor imagery data of stroke patiens.

Installation

1. Create a conda environment

conda create -n env_name python=3.10

2. Install pip package

pip install -r requirements.txt

3. Install seaborn package

pip install seaborn==0.12.0

Dataset

The EEG dataset of stroke patients is provided by Liu et.al in https://doi.org/10.6084/m9.figshare.21679035.v5
You just need to download "sourcedata.zip" through this link and unzip it to the "dataset/sourcedata" directory.

Usage

1. Python file: figshare_stroke_fc2.py

• Plot functional connectivity matrix and corresponding topology in 3 frequency bands for 50 stroke patients.
• Save the functional connectivity data (imcoh_left.npy and imcoh_right.npy) to data_load/ImCoh_data.

python figshare_stroke_fc2.py

2. Python file: figshare_fc_mst2.py

• Calculate and visualize the maximum spanning tree (MST) transformed from the function connectivity matrix.
• Correlation analysis: regplot between the NIHSS score and various MST metrics (diameter, eccentricity, leaf number, tree hierarchy).
• Comparision analysis: violinplot of the MST metrics under the low NIHSS group and high NIHSS group.
• Correct the correlation coefficient by Spearman correlation and permutation test.

python figshare_fc_mst2.py

Directory

│  figshare_fc_mst2.py
│  figshare_stroke_fc2.py
│  
├─dataset
│  │  subject.csv
│  │  
│  └─sourcedata
│      ├─sub-01
│      │      sub-01_task-motor-imagery_eeg.mat
│      │      
│      ├─sub-02
│      │      sub-02_task-motor-imagery_eeg.mat
│      │ 
│      │  ...
│      │ 
│      └─sub-50
│              sub-50_task-motor-imagery_eeg.mat
│              
└─data_load
    └─ImCoh_data
        └─alpha_beta12
                imcoh_left.npy
                imcoh_right.npy