Brain MRI Segmentation Comparison Study

This project focuses on comparing three different brain MRI segmentation methods:

• FreeSurfer (recon-all pipeline)
• SynthSeg
• Brainchop

Project Overview

The project aims to evaluate and compare the reproducibility of brain segmentation and parcellation results across different methods and datasets:

Datasets

1. Simon Dataset
2. Traveling Subjects Dataset (SRPBS)

Analysis Goals

• Compare segmentation reproducibility across methods
• Evaluate parcellation consistency within and between subjects
• Generate comparative visualizations of results
• Assess method-specific strengths and limitations

Structure

brain_mri_segmentation/
├── data/                    # Data directory
│   ├── raw/                # Raw MRI data
│   │   ├── simon/         # Simon dataset
│   │   └── traveling/     # Traveling subjects dataset
│   └── processed/         # Processed and segmented data
│       ├── freesurfer/    # FreeSurfer results
│       ├── synthseg/      # SynthSeg results
│       └── brainchop/     # Brainchop results
├── src/                    # Source code
│   ├── preprocessing/     # Data preprocessing scripts
│   ├── segmentation/      # Segmentation pipeline scripts
│   │   ├── freesurfer/   # FreeSurfer pipeline
│   │   ├── synthseg/     # SynthSeg pipeline
│   │   └── brainchop/    # Brainchop pipeline
│   └── evaluation/       # Evaluation metrics and tools
│       ├── reproducibility/  # Reproducibility analysis
│       └── visualization/    # Result visualization
├── tests/                # Test scripts
└── docs/                # Documentation

Requirements

• FreeSurfer 7.x
• SynthSeg
• Brainchop
• Python 3.8+
• Additional dependencies listed in requirements.txt

Installation

1. Clone the repository:

git clone https://github.com/your-username/brain-mri-segmentation.git
cd brain-mri-segmentation

2. Install dependencies:

pip install -r requirements.txt

Usage

[Usage instructions will be added as the project develops]

Expected Outputs

• Reproducibility metrics for each segmentation method
• Comparative visualizations of segmentation results
• Parcellation consistency analysis across subjects
• Statistical analysis of method performance

License

MIT License

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Authors

Ekaterina Kondrateva, Sandgy Barg