Connectome

Created as part of an university project "Innovationlab Big Data Science" at LMU Munich.

About The Project

Together with the project partner Dr. Boris Rauchmann (Uniklinikum München), the Connectome team researched Prediction and Explanation of Alzheimer Diagnosis based on connectivity matrices utilizing the Brainnetome Atlas. The results of the project include a pipeline for processing connectivity matrices in order to predict and explain a patients` AD status. This pipeline enables the user to automate the training, evaluation and interpretation for several models as well as for several dataset options, e.g. aggregated connectivity matrices, connectivity matrices and graph metrics applied to human brain connectivity data.

Getting Started

Follow the following steps to start the pipeline locally.

Prerequisites

• Install python 3.9.x or above
• From either:
• https://www.python.org/downloads/
• or: https://www.anaconda.com/

Recommended: Create a virtual environment before installing the package. If you choose not to, you can skip to the Installation

1. Open the command prompt
2. Navigate to the project directory or create a new one:

mkdir path/to/python/project

3. Create a virtual environment within the project folder
   Linux:

python3 -m venv connectome_env

Windows:

python -m venv connectome_env

4. Activate virtual environment
   Linux:

source connectome_env/bin/activate

Windows:

connectome\Scripts\activate

5. Go to installation

Installation

• Windows/Linux

  git clone https://github.com/JanaGauss/Connectome.git
  cd Connectome
  pip install .

Usage

This is the beginning to our Connectome Pipeline. To use our pipeline just open https://github.com/JanaGauss/Connectome/blob/e9fa13e33a58fb1470c92e59d77b69730b31bc67/Connectome%20Pipeline.ipynb

Important: The excel sheet must contain the column "ConnID" as an identifier for merging with the matlab files.

To start preprocessing the data for analysis, you need a folder with matlab files of connectivity matrices and an excel sheet with subject information. Afterwards, just specify the folder path and your good to go:

matlab_dir = r"./path/to/matlab/files" # Enter the directory for the matlab files
excel_path = r"./path/to/excel/sheet.xlsx" # Enter the directory for the corresponding excel sheet with xlsx at the end

For further the further steps check out the demo video or read our full documentation here: https://likai97.github.io/Conncetome-Documentation/

Authors

• Leo Schaabner
• Kai Becker
• Jana Gauß
• Katharina Brenner
• Jonas Klingele