Brain Mapper

This repository contains code for spectrum analysis of ADHD and ASD disorders using Mapper Algorithm.

Getting Started

Prerequisites

• Python >= 3.8
• Bash (for running generator.sh)
• Recommended: Create a virtual environment using venv

Install dependencies:

pip install -r requirements.txt

Running the pipeline

Run each step sequentially:

Preprocess data

python3 preprocess.py

• This script performs the following:

  • Loads raw static functional connectivity (SFC) matrices from .mat files for Autism, Asperger, and ADHD subtypes.

  • Aligns ROI indices to a common 190-ROI standard using a provided Excel mapping file, ensuring all datasets have consistent spatial references for further analysis.

Split dataset

python3 split_dataset.py

This script splits the dataset into discovery set and replication set.

Evaluate parameters

python3 evaluate_parameters.py

This script generates plots of DBSCAN epsilon and projection functions to guide the parameter selection.

Generate mapper graphs

After reviewing the plots generated by evaluate_parameters.py, choose the DBSCAN epsilon, minPTS, and projection function that show optimal clustering performance and run the generator script with your selected parameters:

bash generator.sh

Run mapper interactive GUI

cd MapperInteractive
python3 run.py

This step launches the interactive Mapper visualization tool. In the GUI:

1. Load the generated graphs directory in MapperInteractive/app/static/uploads.
2. Choose the graph you want to inspect.
3. Display and explore the graph interactively.

Extract transition features

python3 transition_features.py