TI-CSC Toolbox for Docker Image

Developed and maintained by Ido Haber - [email protected]
Last update: October 21, 2024

---

Overview

The TI-CSC Toolbox is designed for researchers and engineers involved in Temporal Interference (TI) stimulation. This CLI-based toolbox facilitates the optimization of montages and the analysis of unipolar and bipolar montages. All dependencies are preinstalled & configured using docker, to facilitate easy of use and collaboration.

Container Contents

The Docker container includes the following tools and libraries:

• FSL 6.0.6
• Freesurfer 7.4.1
• SimNIBS 4.1.0
• dcm2niix
• MATLAB Runtime r2024a
• Git repository with analysis and optimization scripts
• Commonly used CLI tools: VIM, NVIM, TMUX, Git, and more.

---

Compatibility

• Operating Systems: Linux, Windows, macOS
• Please reach out if you encounter any bugs or issues.

---

How to Run the Docker Container

1. Ensure Docker is Installed:

   • Download and install Docker Desktop for macOS and Windows, or Docker Engine for Linux.

2. Set Up XQuartz or Xming:

   • For macOS: Install XQuartz.
   • For Windows: Install Xming.
   • Note: These are only necessary if you plan to use GUI functionality.
   • macOS tends to be GUI problematic. It is highly recommanded to run the config_sys.sh script before trying the run the containers. Also, consider downgrading Xquartz to an older version.

3. Set Up Project Directory:

   • Ensure your project directory follows this structure:

        ├── MRIs
        │   ├── 001
        │   │   ├── T1
        │   │   └── T2
        │   └── 002
        │       ├── T1
        │       └── T2
        └── Subjects
            ├── sub_001
            └── sub_002
     
     

4. Run the Docker Container:

   • On Unix systems (Linux/macOS), use the provided starter bash script:

     bash compose-loader.sh   # Recommended.
     bash img-loader.sh       # A single large image which contains everything. 

   • On Windows, if you do not have bash available, run the following command manually:

     docker run --rm -ti -e DISPLAY=host.docker.internal:0.0 -v C:\path\to\project_dir:/mnt/project_dir -v "$LOCAL_PROJECT_DIR":/mnt/"$PROJECT_DIR_NAME" idossha/ti-package:vx.x.x

   • Replace C:\path\to\project_dir with the actual path to your project directory on Windows.

---

Tips

• Automatic Screenshots:
  If you wish to enable automatic screenshots in the main scripts, simply uncomment the relevant lines in the script files.

• Clearing Previous Outputs:
  Before re-executing analysis or optimization, it is highly recommended to clear or remove previous outputs to avoid conflicts.

• MATLAB Runtime Warnings:
  MATLAB Runtime may display warnings, but they generally do not affect the correctness of the output.

---

To create head models

1. locate you MR scans

• If these are DICOM, use dcm2niix function from the command line to transform to niftis
• Once you have your niftis in place, use the charm function from SimNIBS to reconstruct the head model and co-register EEG nets.
• Once m2m_subjectrID you can move to the analyzer / optimizer

Analyzer Requirements:

1. A project directory containing the Subjects subdirectory with m2m_SubjectID directories.
2. A tensor file (required only for anisotropic simulation).

How to Run:

bash start-ana.sh

Follow the on-screen prompts.
If running multiple consecutive analyses, move the previous sim_SubjectID directory elsewhere before starting a new one.

Optimizer Requirements:

A project directory containing the Subjects subdirectory with m2m_SubjectID directories.

How to Run:

bash start-opt.sh

Follow the on-screen prompts.
First optimization for every subject will require you to create two leadfields which might take > 1h + Following optimizations will automatically skip this step. For hd-EEG optimization, allocate more RAM to Docker. Recommended: >32GB.

---

Cheers,
Ido Haber