⚙️ Getting Started

1. Install dependencies
   Run the following command in your terminal:

   pip install -r requirements.txt

   If there is an error requiring the install of Rust and Cargo, please run the following command:

   curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

   You might want to delete this from your PATH variables after running the code

2. Place the scan folders

   • Move Methodology1_Scans into the Methodology #1 directory.
   • Move Methodology2_Scans into the Methodology #2 directory.

   Both folders are available in the Canvas assignment titled “Week 13: Final Report and Presentation” Note:This project uses python 3.10

---

🚀 Running the Code

Note: If there is a file not found error, the "path" variable inside the codes can be changed to the proper folders necessary

🔹 Methodology #1

1. Navigate to the Methodology #1 directory.

2. Run the following scripts in order:

3. Note: You do not have to run these commands, they are here to let you know which files to run

   python 'Methodology #1/Visiontransformer.py'
   python 'Methodology #1/XGBoost_ML.py'

🔹 Methodology #2

1. Navigate to the Methodology #2 directory.
2. Run the following scripts in sequence:

   python 'Methodology #2/Create_FC_Dataset.py'
   python 'Methodology #2/FC_NN_Stage1.py'
   python 'Methodology #2/FC_NN_Stage2.py'

---

📁 Folder Structure

OCD_Classification_Project/
│
├── Methodology #1/
│   ├── Methodology1_Scans/
│   ├── Visiontransformer.py
│   └── XGBoost_ML.py
│
├── Methodology #2/
│   ├── Methodology2_Scans/
│   ├── Create_FC_Dataset.py
│   ├── FC_NN_Stage1.py
│   └── FC_NN_Stage2.py
│
└── requirements.txt

---

🧠 Project Overview

This project implements two distinct pipelines for classifying OCD patients based on neuroimaging data:

• Methodology #1: Image-based classification using Vision Transformers and XGBoost.
• Methodology #2: Functional connectivity matrix construction followed by a two-stage neural network classifier.

---

📬 Contact

For questions or collaboration inquiries, please reach out via Canvas or email.