Application of explainability methods to brain age prediction models

This code demo is part of May 2024's Translational Science Methods Club: Multimodal Data Integration

Brain Age Prediction Model

Implementation of a 4-layer neural network designed to predict brain age from synthetic input features. We use the covariance matrix to preserve the relationships between features and target variable Age at Visit during generation of data.

Interpretability Methods

We implement several interpretability methods from the Captum library to examine feature importance.

Setup Instructions

Forking the Repository

Navigate to the GitHub repository and click the "Fork" button at the top-right corner of the page to create a copy of the repository in your GitHub account.

Cloning the Repository

Once you have forked the repository, you need to clone it to your local machine. Open your terminal and run the following command:
git clone https://github.com/<your-username>/TSciM-Club_May2024.git
Replace with your GitHub username.

Setup Environment

cd TSciM-Club_May2024
pip install -r requirements.txt